This project is a Tcl extension for OpenCV library.
OpenCV (Open Source Computer Vision Library) is a library of programming functions mainly aimed at real-time computer vision.
This extension requires OpenCV 4.5.0 or newer.
See BUILDING.md to know how to build the extension.
MATRIX dims
MATRIX size
MATRIX rows
MATRIX cols
MATRIX channels
MATRIX depth
MATRIX type
MATRIX empty
MATRIX at index_list channel ?value?
MATRIX inv ?method?
MATRIX dot matrix
MATRIX cross matrix
MATRIX * matrix
MATRIX add value
MATRIX subtract value
MATRIX multiply value
MATRIX divide value
MATRIX transpose
MATRIX diag ?d?
MATRIX crop x y width height
MATRIX rect x y width height
MATRIX copyTo matrix ?mask?
MATRIX convertTo type ?scale shift?
MATRIX col index
MATRIX colRange startcol endcol
MATRIX row index
MATRIX rowRange startcol endcol
MATRIX pop_back ?nelems?
MATRIX push_back matrix
MATRIX reshape cn ?rows?
MATRIX setData list_data
MATRIX setTo color_list ?mask?
MATRIX toByteArray
MATRIX fromByteArray width height bpp bytes
MATRIX close
::cv::CV_8UC num
::cv::CV_8SC num
::cv::CV_16UC num
::cv::CV_16SC num
::cv::CV_32SC num
::cv::CV_32FC num
::cv::CV_64FC num
::cv::Mat::Mat rows cols type ?color_list?
::cv::Mat::MatWithDims dims size_list type ?color_list?
::cv::Mat::diag matrix
::cv::Mat::eye rows cols type
::cv::Mat::ones rows cols type
::cv::Mat::zeros rows cols type
::cv::matvar name mat
::cv::matvar is provided which does some automatic resource/life-cycle
management using traces on variables. It's an experimental combination of
the set command with a unset/write variable trace on the (local or global)
variable. You can check
example.
::cv::abs matrix
::cv::absdiff matrix_1 matrix_2
::cv::add matrix_1 matrix_2
::cv::addWeighted matrix_1 alpha matrix_2 beta gamma
::cv::bitwise_and matrix_1 matrix_2 ?mask?
::cv::bitwise_or matrix_1 matrix_2 ?mask?
::cv::bitwise_xor matrix_1 matrix_2 ?mask?
::cv::bitwise_not matrix ?mask?
::cv::calcCovarMatrix matrix flags ?ctype?
::cv::cartToPolar matrix_1 matrix_2 ?angleInDegrees?
::cv::compare matrix_1 matrix_2 cmpop
::cv::convertScaleAbs matrix alpha beta
::cv::copyMakeBorder matrix top bottom left right borderType ?color_list?
::cv::countNonZero matrix
::cv::hasNonZero matrix
::cv::determinant matrix
::cv::divide matrix_1 matrix_2 ?scale?
::cv::eigen matrix
::cv::eigenNonSymmetric matrix
::cv::exp matrix
::cv::extractChannel matrix coi
::cv::findNonZero matrix
::cv::flip matrix flipCode
::cv::getOptimalDFTSize vecsize
::cv::dft matrix ?falgs nonzeroRows?
::cv::inRange matrix color_list1 color_list2
::cv::insertChannel src_matrix dst_matrix coi
::cv::log matrix
::cv::LUT matrix1 matrix2
::cv::Mahalanobis matrix1 matrix2 icovar_matrix
::cv::magnitude matrix1 matrix2
::cv::max matrix1 matrix2
::cv::meanStdDev matrix
::cv::min matrix1 matrix2
::cv::minMaxIdx matrix
::cv::minMaxLoc matrix
::cv::multiply matrix_1 matrix_2 ?scale?
::cv::mulSpectrums matrix_1 matrix_2 flags ?conjB?
::cv::split matrix
::cv::merge matrix_list
::cv::norm matrix norm_type
::cv::norm matrix_1 matrix_2 norm_type
::cv::normalize matrix alpha beta norm_type
::cv::pow matrix power
::cv::randn matrix mean_list stddev_list
::cv::randu matrix min_list max_list
::cv::randShuffle matrix ?iterFactor?
::cv::reduce matrix dim rtype ?dtype?
::cv::repeat matrix ny nx
::cv::rotate matrix rotateCode
::cv::setRNGSeed seed
::cv::solve matrix_1 matrix_2 ?flags?
::cv::solveCubic matrix
::cv::solvePoly matrix ?maxIters?
::cv::sortIdx matrix flags
::cv::sqrt matrix
::cv::subtract matrix_1 matrix_2
::cv::sum matrix
::cv::SVBackSubst matrix_w matrix_u matrix_vt matrix
::cv::SVDecomp matrix ?flags?
::cv::trace matrix
::cv::transform matrix_1 matrix_2
::cv::hconcat matrix_1 matrix_2
::cv::vconcat matrix_1 matrix_2
OpenCV 4.8.0 add ::cv::hasNonZero.
::cv::kmeans matrix k bestLabels termCriteria attempts flags
If users setup bestLabels to None, then flags should not set to $::cv::KMEANS_USE_INITIAL_LABELS.
If flags is set to $::cv::KMEANS_USE_INITIAL_LABELS, users should provide a CV_32S matrix.
::cv::perspectiveTransform src_list transformation_matrix
::cv::getBuildInformation
::cv::getTickCount
::cv::getTickFrequency
::cv::PCA matrix flags ?maxComponents?
PCA mean
PCA eigenvalues
PCA eigenvectors
PCA backProject matrix
PCA project matrix
PCA close
::cv::TermCriteria ?type maxCount epsilon?
TermCriteria close
::cv::fromByteArray width height bpp bytes
Creates a new cv::Mat from a gray (bpp equals 1) or RGB (bpp equals 3) byte array with width columns and height rows.
Same rule regarding arguments applies to MATRIX fromByteArray ....
Likewise, the command MATRIX toByteArray returns a 4 element result list.
::cv::FileStorage
FS open filename|data mode
FS keys ?name ...?
FS startMap name
FS startSeq name
FS endMap
FS endSeq
FS readDouble name ...
FS readInt name ...
FS readMat name ...
FS readObj name ...
FS readString name ...
FS writeDouble name double ...
FS writeInt name int ...
FS writeMat name matrix
FS writeObj name obj
FS writeString name string ...
FS close
For more FileStorage info, you can check XML/YAML Persistence.
::cv::imread filename ?flags?
::cv::imdecode bytes ?flags?
::cv::imwrite filename matrix
::cv::imencode fileext matrix
::cv::applyColorMap matrix colormap
::cv::cvtColor src_matrix code ?dstCn?
::cv::calcBackProject matrix channels hist_matrix ranges_list ?scale uniform?
::cv::calcHist matrix channels mask dims histSize_list ranges_list ?uniform accumulate?
::cv::compareHist matrix_1 matrix_2 method
::cv::equalizeHist matrix
::cv::EMD signature1 signature2 distType ?cost?
::cv::floodFill matrix seed_x seed_y color_list ?rect_list loDiff_color upDiff_color flags?
::cv::grabCut matrix x y width height iterCount
::cv::matchTemplate matrix templ_matrix method
::cv::moments matrix ?binaryImage?
::cv::getRotationMatrix2D x y angle scale
::cv::getRectSubPix matrix width height center_x center_y
::cv::HuMoments moments_list
::cv::integral matrix ?sdepth sqdepth?
::cv::remap matrix map1 map2 interpolation
::cv::resize matrix width height ?flags?
::cv::threshold matrix thresh maxval type
::cv::adaptiveThreshold matrix maxValue adaptiveMethod thresholdType blockSize C
::cv::getAffineTransform src_list dst_list
::cv::warpAffine src_matrix transformation_matrix width height ?flags?
::cv::getPerspectiveTransform src_list dst_list ?solveMethod?
::cv::warpPerspective src_matrix transformation_matrix width height ?flags?
::cv::warpPolar matrix dsize_width dsize_height center_x center_y maxRadius ?flags?
::cv::filter2D src_matrix kernel_matrix ?anchor_x anchor_y delta borderType?
::cv::sepFilter2D src_matrix kernelX kernelY ?anchor_x anchor_y delta borderType?
::cv::getDerivKernels dx dy ksize ?normalize ktype?
::cv::getGaborKernel ksize_width ksize_height sigma theta lambd gamma ?psi type?
::cv::getGaussianKernel ksize sigma ?type?
::cv::blur src_matrix ksize_width ksize_height ?anchor_x anchor_y borderType?
::cv::GaussianBlur src_matrix ksize_width ksize_height sigmaX ?sigmaY borderType?
::cv::medianBlur src_matrix ksize
::cv::stackBlur src_matrix ksize_width ksize_height
::cv::bilateralFilter src_matrix d sigmaColor sigmaSpace ?borderType?
::cv::boxFilter src_matrix ksize_width ksize_height ?anchor_x anchor_y normalize borderType?
::cv::sqrBoxFilter src_matrix ksize_width ksize_height ?anchor_x anchor_y normalize borderType?
OpenCV 4.7.0 add ::cv::stackBlur.
::cv::getStructuringElement shape ksize_width ksize_height ?anchor_x anchor_y?
::cv::dilate src_matrix kernel_matrix ?anchor_x anchor_y iterations?
::cv::erode src_matrix kernel_matrix ?anchor_x anchor_y iterations?
::cv::morphologyEx src_matrix op kernel_matrix ?anchor_x anchor_y iterations?
::cv::buildPyramid src_matrix maxlevel ?borderType?
::cv::pyrUp src_matrix ?dstsize_width dstsize_height borderType?
::cv::pyrDown src_matrix ?dstsize_width dstsize_height borderType?
::cv::pyrMeanShiftFiltering src_matrix sp sr ?maxLevel?
::cv::createHanningWindow winSize_width winSize_height type
::cv::phaseCorrelate matrix_1 matrix_2 ?window?
::cv::Canny matrix threshold1 threshold2 ?apertureSize L2gradient?
::cv::Sobel matrix dx dy ?ksize scale delta borderType?
::cv::Scharr matrix dx dy ?scale delta borderType?
::cv::Laplacian matrix ?ksize scale delta borderType?
::cv::distanceTransform matrix distanceType maskSize
::cv::connectedComponents matrix ?connectivity?
::cv::connectedComponentsWithStats matrix ?connectivity?
::cv::watershed matrix markers
::cv::goodFeaturesToTrack matrix maxCorners qualityLevel minDistance ?mask blockSize useHarrisDetector k?
::cv::cornerHarris matrix blockSize ksize k ?borderType?
::cv::cornerEigenValsAndVecs matrix blockSize ksize ?borderType?
::cv::cornerMinEigenVal matrix blockSize ksize ?borderType?
::cv::cornerSubPix matrix corners winSize_width winSize_height zeroZone_widht zeroZone_height termCriteria
::cv::HoughCircles matrix method dp minDist ?param1 param2 minRadius maxRadius?
::cv::HoughLines matrix rho theta threshold ?srn stn min_theta max_theta?
::cv::HoughLinesP matrix rho theta threshold ?minLineLength maxLineGap?
::cv::findContours matrix mode method ?offset_point_x offset_point_y?
::cv::findContoursWithHierarchy matrix mode method ?offset_point_x offset_point_y?
::cv::drawContours matrix contours_list contourIdx color_list thickness ?lineType maxLevel offset_point_x offset_point_y?
::cv::drawContoursWithHierarchy matrix contours_list contourIdx color_list thickness lineType hierarchy maxLevel ?offset_point_x offset_point_y?
::cv::approxPolyDP contour epsilon closed
::cv::arcLength contour closed
::cv::contourArea contour ?oriented?
::cv::boundingRect contour
::cv::minAreaRect contour
::cv::fitEllipse contour
::cv::fitLine contour distType param reps aeps
::cv::boxPoints contour
::cv::minEnclosingCircle contour
::cv::convexHull contour ?clockwise returnPoints?
::cv::convexityDefects contour convexhull
::cv::matchShapes contour1 contour2 method
::cv::pointPolygonTest contour x y measureDist
If users want to find the convexity defects of a contour,
convex hull obtained using ::cv::convexHull that required to set returnPoints to 0.
If users need to know contour's moments, users need to use ::cv::Mat::Mat to create
a matrix and fill the contour's data to the matrix, then use ::cv::moments to calculate.
::cv::arrowedLine matrix point_x1 point_y1 point_x2 point_y2 color_list thickness ?lineType shift tipLength?
::cv::circle matrix center_x center_y radius color_list thickness ?lineType shift?
::cv::clipLine size_list point1_list point2_list
::cv::drawMarker matrix point_x point_y color_list ?markerType markerSize thickness line_type?
::cv::ellipse matrix center_x center_y width hgieht angle startAngle endAngle color_list thickness ?lineType shift?
::cv::fillConvexPoly matrix point_list color_list ?lineType shift?
::cv::fillPoly matrix point_lists color_list ?lineType shift offset_x offset_y?
::cv::getFontScaleFromHeight fontFace pixelHeight ?thickness?
::cv::getTextSize text fontFace fontScale thickness
::cv::line matrix point_x1 point_y1 point_x2 point_y2 color_list thickness ?lineType shift?
::cv::polylines matrix point_list ncontours isClosed color_list thickness ?lineType shift?
::cv::putText matrix text point_x point_y fontFace fontScale color_list thickness ?lineType bottomLeftOrigin?
::cv::rectangle matrix point_x1 point_y1 point_x2 point_y2 color_list thickness ?lineType shift?
::cv::CLAHE ?clipLimit tileGridSize_w tileGridSize_h?
CLAHE apply matrix
CLAHE close
Please notice, CLAHE command will only have 1 instance.
::cv::GeneralizedHoughBallard
GeneralizedHoughBallard detect matrix
GeneralizedHoughBallard getCannyHighThresh
GeneralizedHoughBallard getCannyLowThresh
GeneralizedHoughBallard getDp
GeneralizedHoughBallard getLevels
GeneralizedHoughBallard getMinDist
GeneralizedHoughBallard getVotesThreshold
GeneralizedHoughBallard setCannyHighThresh
GeneralizedHoughBallard setCannyLowThresh value
GeneralizedHoughBallard setDp value
GeneralizedHoughBallard setLevels value
GeneralizedHoughBallard setMinDist value
GeneralizedHoughBallard setTemplate matrix
GeneralizedHoughBallard setVotesThreshold value
GeneralizedHoughBallard close
Please notice, GeneralizedHoughBallard command will only have 1 instance.
::cv::GeneralizedHoughGuil
GeneralizedHoughGuil detect matrix
GeneralizedHoughGuil getAngleEpsilon
GeneralizedHoughGuil getAngleStep
GeneralizedHoughGuil getAngleThresh
GeneralizedHoughGuil getCannyHighThresh
GeneralizedHoughGuil getCannyLowThresh
GeneralizedHoughGuil getDp
GeneralizedHoughGuil getLevels
GeneralizedHoughGuil getMaxAngle
GeneralizedHoughGuil getMaxScale
GeneralizedHoughGuil getMinDist
GeneralizedHoughGuil getMinAngle
GeneralizedHoughGuil getMinScale
GeneralizedHoughGuil getPosThresh
GeneralizedHoughGuil getScaleStep
GeneralizedHoughGuil getScaleThresh
GeneralizedHoughGuil getXi
GeneralizedHoughGuil setAngleEpsilon value
GeneralizedHoughGuil setAngleStep value
GeneralizedHoughGuil setAngleThresh value
GeneralizedHoughGuil setCannyHighThresh value
GeneralizedHoughGuil setCannyLowThresh value
GeneralizedHoughGuil setDp value
GeneralizedHoughGuil setLevels value
GeneralizedHoughGuil setMaxAngle value
GeneralizedHoughGuil setMaxScale value
GeneralizedHoughGuil setMinDist value
GeneralizedHoughGuil setMinAngle value
GeneralizedHoughGuil setMinScale value
GeneralizedHoughGuil setPosThresh value
GeneralizedHoughGuil setScaleStep value
GeneralizedHoughGuil setScaleThresh value
GeneralizedHoughGuil setXi value
GeneralizedHoughGuil setTemplate matrix
GeneralizedHoughGuil close
Please notice, GeneralizedHoughGuil command will only have 1 instance.
::cv::LineSegmentDetector ?refine scale sigma_scale quant ang_th log_eps density_th n_bins?
LineSegmentDetector detect matrix
LineSegmentDetector drawSegments matrix lines
LineSegmentDetector close
LineSegmentDetector's implementation has been removed from OpenCV version 3.4.6 to 3.4.15 and version 4.1.0 to 4.5.3 due original code license conflict. Restored in 4.5.4 again after Computation of a NFA code published under the MIT licens.
Please notice, LineSegmentDetector command will only have 1 instance.
::cv::segmentation::IntelligentScissorsMB
IntelligentScissorsMB applyImage matrix
IntelligentScissorsMB buildMap x y
IntelligentScissorsMB getContour x y ?backward?
IntelligentScissorsMB setEdgeFeatureCannyParameters threshold1 threshold2 ?apertureSize L2gradient?
IntelligentScissorsMB setEdgeFeatureZeroCrossingParameters gradient_magnitude_min_value
IntelligentScissorsMB setGradientMagnitudeMaxLimit gradient_magnitude_threshold_max
IntelligentScissorsMB setWeights weight_non_edge weight_gradient_direction weight_gradient_magnitude
IntelligentScissorsMB close
OpenCV 4.5.2 add ::cv::segmentation::IntelligentScissorsMB.
It is used to find the path (contour) between two points which can be used for image segmentation.
::cv::VideoCapture file/index filename/number ?flags? ?paramId value?
VideoCapture isOpened
VideoCapture read
VideoCapture get propId
VideoCapture set propId value
VideoCapture close
::cv::VideoCapture is using to open video file or open a camera for video
capturing. Second argument is file or index, then the third arugment
is filename or camera id.
::cv::VideoWriter filename fourcc fps width height ?isColor?
VideoWriter isOpened
VideoWriter write matrix
VideoWriter get propId
VideoWriter set propId value
VideoWriter close
::cv::namedWindow winname ?flags?
::cv::imshow winname matrix
::cv::waitKey delay
::cv::moveWindow winname x y
::cv::resizeWindow winname width height
::cv::setWindowTitle winname title
::cv::destroyWindow winname
::cv::destroyAllWindows
::cv::selectROI matrix ?showCrosshair fromCenter?
::cv::setMouseCallback winname callback_code
::cv::createTrackbar trackbarname winname init_value max_value callback_code
::cv::getTrackbarPos trackbarname winname
::cv::drawKeypoints matrix keypoints dst_image color_list ?flags?
::cv::drawMatches matrix1 keypoints1 matrix2 keypoints2 matches1to2 dst_image matchColor_list singlePointColor_list ?flags?
If users setup dst_image to None, then flags should not set to $::cv::DRAW_MATCHES_FLAGS_DRAW_OVER_OUTIMG.
If flags is set to $::cv::DRAW_MATCHES_FLAGS_DRAW_OVER_OUTIMG, users should provide a dst matrix.
::cv::FastFeatureDetector ?threshold nonmaxSuppression type?
FastFeatureDetector detect matrix
FastFeatureDetector getNonmaxSuppression
FastFeatureDetector getThreshold
FastFeatureDetector getType
FastFeatureDetector setNonmaxSuppression value
FastFeatureDetector setThreshold value
FastFeatureDetector setType value
FastFeatureDetector close
Please notice, FastFeatureDetector command will only have 1 instance.
::cv::AgastFeatureDetector ?threshold nonmaxSuppression type?
AgastFeatureDetector detect matrix
AgastFeatureDetector getNonmaxSuppression
AgastFeatureDetector getThreshold
AgastFeatureDetector getType
AgastFeatureDetector setNonmaxSuppression value
AgastFeatureDetector setThreshold value
AgastFeatureDetector setType value
AgastFeatureDetector close
Please notice, AgastFeatureDetector command will only have 1 instance.
::cv::MSER ?delta min_area max_area max_variation min_diversity max_evolution area_threshold min_margin edge_blur_size?
MSER detectRegions matrix
MSER getDelta
MSER getMaxArea
MSER getMinArea
MSER getPass2Only
MSER setDelta value
MSER setMaxArea value
MSER setMinArea value
MSER setPass2Only value
MSER close
Please notice, MSER command will only have 1 instance.
::cv::ORB ?nfeatures scaleFactor nlevels edgeThreshold firstLevel WTA_K scoreType patchSize fastThreshold?
ORB detect matrix
ORB compute matrix keypoints
ORB detectAndCompute matrix
ORB getEdgeThreshold
ORB getFastThreshold
ORB getFirstLevel
ORB getMaxFeatures
ORB getNLevels
ORB getPatchSize
ORB getScaleFactor
ORB getScoreType
ORB getWTA_K
ORB setEdgeThreshold value
ORB setFastThreshold value
ORB setFirstLevel value
ORB setMaxFeatures value
ORB setNLevels value
ORB setPatchSize value
ORB setScaleFactor value
ORB setScoreType value
ORB setWTA_K value
ORB close
Please notice, ORB command will only have 1 instance.
::cv::AKAZE ?descriptor_type descriptor_size descriptor_channels threshold nOctaves nOctaveLayers diffusivity?
AKAZE detect matrix
AKAZE compute matrix keypoints
AKAZE detectAndCompute matrix
AKAZE getDescriptorChannels
AKAZE getDescriptorSize
AKAZE getDescriptorType
AKAZE getDiffusivity
AKAZE getNOctaveLayers
AKAZE getNOctaves
AKAZE getThreshold
AKAZE setDescriptorChannels value
AKAZE setDescriptorSize value
AKAZE setDescriptorType value
AKAZE setDiffusivity value
AKAZE setNOctaveLayers value
AKAZE setNOctaves value
AKAZE setThreshold value
AKAZE close
Please notice, AKAZE command will only have 1 instance.
::cv::BRISK ?thresh octaves patternScale?
BRISK detect matrix
BRISK compute matrix keypoints
BRISK detectAndCompute matrix
BRISK close
Please notice, BRISK command will only have 1 instance.
::cv::KAZE ?extended upright threshold nOctaves nOctaveLayers diffusivity?
KAZE detect matrix
KAZE compute matrix keypoints
KAZE detectAndCompute matrix
KAZE getDiffusivity
KAZE getExtended
KAZE getNOctaveLayers
KAZE getNOctaves
KAZE getThreshold
KAZE getUpright
KAZE setDiffusivity value
KAZE setExtended value
KAZE setNOctaveLayers value
KAZE setNOctaves value
KAZE setThreshold value
KAZE setUpright value
KAZE close
Please notice, KAZE command will only have 1 instance.
::cv::SIFT ?nfeatures nOctaveLayers contrastThreshold edgeThreshold sigma?
SIFT detect matrix
SIFT compute matrix keypoints
SIFT detectAndCompute matrix
SIFT close
SIFT (Scale-Invariant Feature Transform) algorithm has been moved to the OpenCV main repository in OpenCV 4.4.0 because its patent expired.
Please notice, SIFT command will only have 1 instance.
::cv::AffineFeature backend ?maxTilt minTilt tiltStep rotateStepBase?
AffineFeature detect matrix
AffineFeature compute matrix keypoints
AffineFeature detectAndCompute matrix
AffineFeature close
OpenCV 4.5.0 add AffineFeature (A-SIFT) API.
Please notice, AffineFeature command will only have 1 instance. And backend supports SIFT, KAZE, ORB, AKAZE and BRISK.
::cv::BFMatcher normType crossCheck
BFMatcher match queryDescriptors trainDescriptors
BFMatcher knnMatch queryDescriptors trainDescriptors k
BFMatcher close
Please notice, BFMatcher command will only have 1 instance.
::cv::FlannBasedMatcher ?algorithm indexParams?
FlannBasedMatcher match queryDescriptors trainDescriptors
FlannBasedMatcher knnMatch queryDescriptors trainDescriptors k
FlannBasedMatcher close
algorithm can specify FLANN_INDEX_LSH or FLANN_INDEX_KDTREE. Default is FLANN_INDEX_LSH.
indexParams is a list of LshIndexParams parameters (table_number, key_size, multi_probe_level) or
KDTreeIndexParams parameters (trees).
Please notice, FlannBasedMatcher command will only have 1 instance.
::cv::SimpleBlobDetector ?-minThreshold value? ?-maxThreshold value? ?-filterByArea boolean? ?-minArea value? ?-maxArea value? ?-filterByCircularity boolean? ?-minCircularity value? ?-maxCircularity value? ?-filterByConvexity boolean? ?-minConvexity value? ?-maxConvexity value? ?-filterByInertia boolean? ?-minInertiaRatio value? ?-maxInertiaRatio value?
SimpleBlobDetector detect matrix
SimpleBlobDetector close
Please notice, SimpleBlobDetector command will only have 1 instance.
::cv::BOWKMeansTrainer clusterCount termCriteria ?attempts flags?
BOWKMeansTrainer add descriptors
BOWKMeansTrainer clear
BOWKMeansTrainer cluster
BOWKMeansTrainer descriptorsCount
BOWKMeansTrainer getDescriptors
BOWKMeansTrainer close
::cv::BOWImgDescriptorExtractor dextractor dmatcher
BOWImgDescriptorExtractor compute matrix keypoints
BOWImgDescriptorExtractor descriptorSize
BOWImgDescriptorExtractor descriptorType
BOWImgDescriptorExtractor getVocabulary
BOWImgDescriptorExtractor setVocabulary vocabulary
BOWImgDescriptorExtractor close
::cv::BOWImgDescriptorExtractor now supports SIFT or KAZE (dextractor) and FlannBasedMatcher (dmatcher).
::cv::findChessboardCorners image patternSize_width patternSize_height ?flags?
::cv::drawChessboardCorners image patternSize_width patternSize_height corners patternWasFound
::cv::calibrateCamera objectPoints_list imagePoints_list width height cameraMatrix distCoeffs
::cv::getOptimalNewCameraMatrix cameraMatrix distCoeffs width height alpha nwidth nheight ?centerPrincipalPoint?
::cv::undistort matrix cameraMatrix distCoeffs newCameraMatrix
::cv::initUndistortRectifyMap cameraMatrix distCoeffs R newCameraMatrix width height m1type
::cv::projectPoints objectPoints rvec tvec cameraMatrix distCoeffs
::cv::solvePnP objectPoints imagePoints cameraMatrix distCoeffs
::cv::findChessboardCorners returns a list of (retval, corners).
The retval value is a non-zero value if all of the corners are found and
they are placed in a certain order.
And the retval value should be passed to ::cv::drawChessboardCorners's
parameter patternWasFound.
::cv::computeCorrespondEpilines matrix whichImage F
::cv::estimateAffine2D matrix_1 matrix_2 ?method ransacReprojThreshold maxIters confidence refineIters?
::cv::estimateAffinePartial2D matrix_1 matrix_2 ?method ransacReprojThreshold maxIters confidence refineIters?
::cv::estimateAffine3D matrix_1 matrix_2 ?ransacThreshold confidence?
::cv::findFundamentalMat matrix_1 matrix_2 ?method ransacReprojThreshold confidence?
::cv::findHomography matrix_1 matrix_2 ?method ransacReprojThreshold maxIters confidence?
::cv::estimateAffine2D, ::cv::estimateAffinePartial2D, ::cv::estimateAffine3D,
::cv::findFundamentalMat and ::cv::findHomography returns a list of (result, inliers or mask).
Note that whenever an H matrix cannot be estimated, an empty one will be returned.
You can use MATRIX empty to check this.
::cv::StereoBM ?numDisparities blockSize?
StereoBM compute matrix1 matrix2
StereoBM getPreFilterCap
StereoBM getPreFilterSize
StereoBM getPreFilterType
StereoBM getSmallerBlockSize
StereoBM getTextureThreshold
StereoBM getUniquenessRatio
StereoBM setPreFilterCap value
StereoBM setPreFilterSize value
StereoBM setPreFilterType value
StereoBM setSmallerBlockSize value
StereoBM setTextureThreshold value
StereoBM setUniquenessRatio value
StereoBM close
Please notice, StereoBM command will only have 1 instance.
::cv::StereoSGBM ?minDisparity numDisparities blockSize P1 P2 disp12MaxDiff preFilterCap uniquenessRatio speckleWindowSize speckleRange mode?
StereoSGBM compute matrix1 matrix2
StereoSGBM getMode
StereoSGBM getP1
StereoSGBM getP2
StereoSGBM getPreFilterCap
StereoSGBM getUniquenessRatio
StereoSGBM setMode value
StereoSGBM setP1 value
StereoSGBM setP2 value
StereoSGBM setPreFilterCap value
StereoSGBM setUniquenessRatio value
StereoSGBM close
Please notice, StereoSGBM command will only have 1 instance.
::cv::BackgroundSubtractorKNN ?history dist2Threshold detectShadows?
BackgroundSubtractorKNN apply matrix
BackgroundSubtractorKNN getDetectShadows
BackgroundSubtractorKNN getDist2Threshold
BackgroundSubtractorKNN getHistory
BackgroundSubtractorKNN setDetectShadows value
BackgroundSubtractorKNN setDist2Threshold value
BackgroundSubtractorKNN setHistory value
BackgroundSubtractorKNN close
Please notice, BackgroundSubtractorKNN command will only have 1 instance.
::cv::BackgroundSubtractorMOG2 ?history varThreshold detectShadows?
BackgroundSubtractorMOG2 apply matrix
BackgroundSubtractorMOG2 getDetectShadows
BackgroundSubtractorMOG2 getHistory
BackgroundSubtractorMOG2 getVarThreshold
BackgroundSubtractorMOG2 setDetectShadows value
BackgroundSubtractorMOG2 setHistory value
BackgroundSubtractorMOG2 setVarThreshold value
BackgroundSubtractorMOG2 close
Please notice, BackgroundSubtractorMOG2 command will only have 1 instance.
::cv::meanShift matrix x y width height termCriteria
::cv::CamShift matrix x y width height termCriteria
::cv::calcOpticalFlowPyrLK prevImg nextImg prevPts winSize_width winSize_height maxLevel termCriteria
::cv::calcOpticalFlowFarneback prevImg nextImg pyr_scale levels winsize iterations poly_n poly_sigma flags
::cv::readOpticalFlow path
::cv::writeOpticalFlow path flow_matrix
::cv::readOpticalFlow and ::cv::writeOpticalFlow in OpenCV 3.x is in
contrib module (optflow). If you build this extension with OpenCV 3.x failed,
you need build optflow module or disable 2 commands by youself.
::cv::computeECC matrix_1 matrix_2 ?mask?
::cv::findTransformECC matrix_1 matrix_2 warpMatrix motionType ?termCriteria maks gaussFiltSize?
Users should use ::cv::Mat::eye to create a 3x3 (for $::Cv::MOTION_HOMOGRAPHY)
or 2x3 warpMatrix then pass to ::cv::findTransformECC.
cv::TrackerMIL
TrackerMIL init matrix x y width height
TrackerMIL update matrix
TrackerMIL close
OpenCV 4.5.1 add cv::TrackerMIL.
Please notice, TrackerMIL command will only have 1 instance.
cv::TrackerGOTURN ?modelBin modelTxt?
TrackerGOTURN init matrix x y width height
TrackerGOTURN update matrix
TrackerGOTURN close
OpenCV 4.5.1 add cv::TrackerGOTURN.
You can get related files from OpenCV extra or
here.
If uesrs need more info, I think you can check OpenCV documentation.
Please notice, TrackerGOTURN command will only have 1 instance.
cv::TrackerDaSiamRPN ?-model value? ?-kernel_cls1 value? ?-kernel_r1 value? ?-backend value? ?-target value?
TrackerDaSiamRPN init matrix x y width height
TrackerDaSiamRPN update matrix
TrackerDaSiamRPN getTrackingScore
TrackerDaSiamRPN close
OpenCV 4.5.3 add cv::TrackerDaSiamRPN. You can get links to onnx models from
C++ sample.
Please notice, TrackerDaSiamRPN command will only have 1 instance.
::cv::inpaint matrix inpaintMask inpaintRadius flags
::cv::decolor matrix
::cv::fastNlMeansDenoising matrix ?h templateWindowSize searchWindowSize?
::cv::fastNlMeansDenoisingColored matrix ?h hColor templateWindowSize searchWindowSize?
::cv::fastNlMeansDenoising expected to be applied to grayscale images.
::cv::fastNlMeansDenoisingColored is for colored images.
::cv::colorChange matrix mask red_mul green_mul blue_mul
::cv::illuminationChange matrix mask alpha beta
::cv::textureFlattening matrix mask low_threshold high_threshold ?kernel_size?
::cv::seamlessClone src_matrix dst_matrix mask x y flags
::cv::detailEnhance matrix ?sigma_s sigma_r?
::cv::edgePreservingFilter matrix ?flags sigma_s sigma_r?
::cv::pencilSketch matrix ?sigma_s sigma_r shade_factor?
::cv::stylization matrix ?sigma_s sigma_r?
Below are High Dynamic Range Imaging related commands:
::cv::AlignMTB ?max_bits exclude_range cut?
AlignMTB process matrix_list
AlignMTB close
Please notice, AlignMTB command will only have 1 instance.
::cv::CalibrateDebevec ?samples lambda random?
CalibrateDebevec process matrix_list times_list
CalibrateDebevec close
Please notice, CalibrateDebevec command will only have 1 instance.
::cv::MergeDebevec
MergeDebevec process matrix_list times_list response
MergeDebevec close
Please notice, MergeDebevec command will only have 1 instance.
::cv::MergeMertens ?contrast_weight saturation_weight exposure_weight?
MergeMertens process matrix_list
MergeMertens close
Please notice, MergeMertens command will only have 1 instance.
::cv::TonemapDrago ?gamma saturation bias?
TonemapDrago process hdrDebevec
TonemapDrago close
Please notice, TonemapDrago command will only have 1 instance.
::cv::TonemapMantiuk ?gamma scale saturation?
TonemapMantiuk process hdrDebevec
TonemapMantiuk close
Please notice, TonemapMantiuk command will only have 1 instance.
::cv::TonemapReinhard ?gamma intensity light_adapt color_adapt?
TonemapReinhard process hdrDebevec
TonemapReinhard close
Please notice, TonemapReinhard command will only have 1 instance.
::cv::Stitcher mode
Stitcher stitch image_list
Stitcher close
Please notice, Stitcher command will only have 1 instance.
::cv::CascadeClassifier filename
CascadeClassifier detectMultiScale matrix ?scaleFactor minNeighbors minWidth minHeight maxWidth maxHeight?
CascadeClassifier close
::cv::HOGDescriptor winSize_width winSize_height blockSize_width blockStride_width blockStride_height blockSize_height cellSize_width cellSize_height nbins ?derivAperture winSigma L2HysThreshold gammaCorrection nlevels signedGradient?
HOGDescriptor compute matrix ?winStride_width winStride_heigth padding_width padding_height?
HOGDescriptor detectMultiScale matrix ?hitThreshold winStride_width winStride_height padding_width padding_height scale finalThreshold useMeanshiftGrouping?
HOGDescriptor getDefaultPeopleDetector
HOGDescriptor getDaimlerPeopleDetector
HOGDescriptor setSVMDetector svmdetector
HOGDescriptor close
:cv::QRCodeDetector
QRCodeDetector detect matrix
QRCodeDetector detectAndDecode matrix
QRCodeDetector close
Please notice, QRCodeDetector looks like have issues if encoded QR Code data is long when I test it.
::cv::FaceDetectorYN model config width height ?score_threshold nms_threshold top_k backend_id target_id?
FaceDetectorYN detect matrix
FaceDetectorYN close
OpenCV 4.5.4 add ::cv::FaceDetectorYN (DNN-based face detector).
Please notice, FaceDetectorYN command will only have 1 instance.
::cv::FaceRecognizerSF model config ?backend_id target_id?
FaceRecognizerSF alignCrop matrix face_box
FaceRecognizerSF feature aligned_img
FaceRecognizerSF match face_feature1 face_feature2 ?dis_type?
FaceRecognizerSF close
OpenCV 4.5.4 add ::cv::FaceRecognizerSF (DNN-based face recognizer).
Please notice, FaceRecognizerSF command will only have 1 instance.
You can check related tutorial and get model download link.
::cv::QRCodeEncoder ?-correction_level value? ?-mode value? ?-structure_number value? ?-version value?
QRCodeEncoder encode string
QRCodeEncoder close
OpenCV 4.5.5 add ::cv::QRCodeEncoder.
Please notice, QRCodeEncoder command will only have 1 instance.
::cv::BarcodeDetector
BarcodeDetector detectAndDecodeWithType matrix
BarcodeDetector close
OpenCV 4.8.0 add ::cv::BarcodeDetector.
Please notice, BarcodeDetector command will only have 1 instance.
::cv::ml::LogisticRegression
::cv::ml::LogisticRegression::load filename
LogisticRegression get_learnt_thetas
LogisticRegression getIterations
LogisticRegression getLearningRate
LogisticRegression getMiniBatchSize
LogisticRegression getRegularization
LogisticRegression getTrainMethod
LogisticRegression setIterations value
LogisticRegression setLearningRate value
LogisticRegression setMiniBatchSize value
LogisticRegression setRegularization value
LogisticRegression setTrainMethod value
LogisticRegression setTermCriteria termCriteria
LogisticRegression train trainData ?flags?
LogisticRegression predict samples ?flags?
LogisticRegression save filename
LogisticRegression close
Please notice, LogisticRegression and LogisticRegression::load command will only have 1 instance.
::cv::ml::NormalBayesClassifier
::cv::ml::NormalBayesClassifier::load filename
NormalBayesClassifier train trainData ?flags?
NormalBayesClassifier predict samples ?flags?
NormalBayesClassifier predictProb samples ?flags?
NormalBayesClassifier save filename
NormalBayesClassifier close
Please notice, NormalBayesClassifier and NormalBayesClassifier::load command will only have 1 instance.
::cv::ml::KNearest
::cv::ml::KNearest::load filename
KNearest getAlgorithmType
KNearest getDefaultK
KNearest getEmax
KNearest getIsClassifier
KNearest setAlgorithmType value
KNearest setDefaultK value
KNearest setEmax
KNearest setIsClassifier
KNearest findNearest samples k
KNearest train trainData ?flags?
KNearest predict samples ?flags?
KNearest save filename
KNearest close
Please notice, KNearest and KNearest::load command will only have 1 instance.
::cv::ml::SVM
::cv::ml::SVM::load filename
SVM getC
SVM getCoef0
SVM getDegree
SVM getGamma
SVM getNu
SVM getP
SVM getType
SVM getKernelType
SVM getDecisionFunction
SVM getSupportVectors
SVM getUncompressedSupportVectors
SVM setC value
SVM setCoef0 value
SVM setDegree value
SVM setGamma value
SVM setNu value
SVM setP value
SVM setType value
SVM setKernel value
SVM setTermCriteria termCriteria
SVM train trainData ?flags?
SVM predict samples ?flags?
SVM save filename
SVM close
Please notice, SVM and SVM::load command will only have 1 instance.
::cv::ml::SVMSGD
::cv::ml::SVMSGD::load filename
SVMSGD getInitialStepSize
SVMSGD getMarginRegularization
SVMSGD getMarginType
SVMSGD getShift
SVMSGD getStepDecreasingPower
SVMSGD getSvmsgdType
SVMSGD getWeights
SVMSGD setInitialStepSize value
SVMSGD setMarginRegularization value
SVMSGD setMarginType value
SVMSGD setOptimalParameters svmsgdType marginType
SVMSGD setStepDecreasingPower value
SVMSGD setSvmsgdType value
SVMSGD setTermCriteria termCriteria
SVMSGD train trainData ?flags?
SVMSGD predict samples ?flags?
SVMSGD save filename
SVMSGD close
Please notice, SVMSGD and SVMSGD::load command will only have 1 instance.
::cv::ml::DTrees
::cv::ml::DTrees::load filename
DTrees getCVFolds
DTrees getMaxCategories
DTrees getMaxDepth
DTrees getMinSampleCount
DTrees getPriors
DTrees getRegressionAccuracy
DTrees getTruncatePrunedTree
DTrees getUse1SERule
DTrees getUseSurrogates
DTrees setCVFolds value
DTrees setMaxCategories value
DTrees setMaxDepth value
DTrees setMinSampleCount value
DTrees setPriors matrix
DTrees setRegressionAccuracy value
DTrees setTruncatePrunedTree value
DTrees setUse1SERule value
DTrees setUseSurrogates value
DTrees train trainData ?flags?
DTrees predict samples ?flags?
DTrees save filename
DTrees close
Please notice, DTrees and DTrees::load command will only have 1 instance.
::cv::ml::Boost
::cv::ml::Boost::load filename
Boost getCVFolds
Boost getMaxCategories
Boost getMaxDepth
Boost getMinSampleCount
Boost getPriors
Boost getRegressionAccuracy
Boost getTruncatePrunedTree
Boost getUse1SERule
Boost getUseSurrogates
Boost getBoostType
Boost getWeakCount
Boost getWeightTrimRate
Boost setCVFolds value
Boost setMaxCategories value
Boost setMaxDepth value
Boost setMinSampleCount value
Boost setPriors matrix
Boost setRegressionAccuracy value
Boost setTruncatePrunedTree value
Boost setUse1SERule value
Boost setUseSurrogates value
Boost setBoostType value
Boost setWeakCount value
Boost setWeightTrimRate value
Boost train trainData ?flags?
Boost predict samples ?flags?
Boost save filename
Boost close
Please notice, Boost and Boost::load command will only have 1 instance.
::cv::ml::RTrees
::cv::ml::RTrees::load filename
RTrees getCVFolds
RTrees getMaxCategories
RTrees getMaxDepth
RTrees getMinSampleCount
RTrees getPriors
RTrees getRegressionAccuracy
RTrees getTruncatePrunedTree
RTrees getUse1SERule
RTrees getUseSurrogates
RTrees getActiveVarCount
RTrees getCalculateVarImportance
RTrees getVarImportance
RTrees getVotes
RTrees setCVFolds value
RTrees setMaxCategories value
RTrees setMaxDepth value
RTrees setMinSampleCount value
RTrees setPriors matrix
RTrees setRegressionAccuracy value
RTrees setTruncatePrunedTree value
RTrees setUse1SERule value
RTrees setUseSurrogates value
RTrees setActiveVarCount value
RTrees setCalculateVarImportance value
RTrees setTermCriteria termCriteria
RTrees train trainData ?flags?
RTrees predict samples ?flags?
RTrees save filename
RTrees close
Please notice, RTrees and RTrees::load command will only have 1 instance.
::cv::ml::ANN_MLP
::cv::ml::ANN_MLP::load filename
ANN_MLP getAnnealCoolingRatio
ANN_MLP getAnnealFinalT
ANN_MLP getAnnealInitialT
ANN_MLP getAnnealItePerStep
ANN_MLP getBackpropMomentumScale
ANN_MLP getBackpropWeightScale
ANN_MLP getRpropDW0
ANN_MLP getRpropDWMax
ANN_MLP getRpropDWMin
ANN_MLP getRpropDWMinus
ANN_MLP getRpropDWPlus
ANN_MLP getLayerSizes
ANN_MLP getTrainMethod
ANN_MLP getWeights layerIdx
ANN_MLP setAnnealCoolingRatio value
ANN_MLP setAnnealFinalT value
ANN_MLP setAnnealInitialT value
ANN_MLP setAnnealItePerStep value
ANN_MLP setBackpropMomentumScale value
ANN_MLP setBackpropWeightScale value
ANN_MLP setRpropDW0 value
ANN_MLP setRpropDWMax value
ANN_MLP setRpropDWMin value
ANN_MLP setRpropDWMinus value
ANN_MLP setRpropDWPlus value
ANN_MLP setActivationFunction value ?param1 param2?
ANN_MLP setLayerSizes matrix
ANN_MLP setTrainMethod value ?param1 param2?
ANN_MLP setTermCriteria termCriteria
ANN_MLP train trainData ?flags?
ANN_MLP predict samples ?flags?
ANN_MLP save filename
ANN_MLP close
Please notice, ANN_MLP and ANN_MLP::load command will only have 1 instance.
::cv::ml::TrainData samples layout responses
TrainData getTestResponses
TrainData getTestSamples
TrainData setTrainTestSplit count ?shuffle?
TrainData setTrainTestSplitRatio ratio ?shuffle?
TrainData close
Please notice, TrainData command will only have 1 instance.
::cv::dnn::blobFromImage matrix scalefactor width height mean_color_list swapRB crop
::cv::dnn::readNet model ?config framework?
::cv::dnn::NMSBoxes bboxes scores score_threshold nms_threshold ?eta top_k?
::cv::dnn::softNMSBoxes bboxes scores score_threshold nms_threshold ?top_k sigma method?
READNET getLayerNames
READNET getUnconnectedOutLayers
READNET getUnconnectedOutLayersNames
READNET setPreferableBackend backendId
READNET setPreferableTarget targetId
READNET setInput blob ?name scalefactor mean_color_list?
READNET forward ?name?
READNET forwardWithNames outBlobNames
READNET close
OpenCV 4.5.4 add ::cv::dnn::softNMSBoxes.
If you want to know supported framework, you can check Deep Learning in OpenCV.
Users could check official OpenCV documentation to know about the different frameworks, their model files and the configuration files.
cv::dnn::TextDetectionModel_EAST model ?config?
TextDetectionModel_EAST detect matrix
TextDetectionModel_EAST getConfidenceThreshold
TextDetectionModel_EAST getNMSThreshold
TextDetectionModel_EAST setConfidenceThreshold value
TextDetectionModel_EAST setNMSThreshold value
TextDetectionModel_EAST setInputParams scalefactor width height mean_color_list swapRB ?crop?
TextDetectionModel_EAST close
TextDetectionModel_EAST is the high-level API for text detection DL networks compatible with EAST model. OpenCV 4.5.1 add the high-level API.
cv::dnn::TextDetectionModel_DB model ?config?
TextDetectionModel_DB detect matrix
TextDetectionModel_DB getBinaryThreshold
TextDetectionModel_DB getMaxCandidates
TextDetectionModel_DB getPolygonThreshold
TextDetectionModel_DB getUnclipRatio
TextDetectionModel_DB setBinaryThreshold value
TextDetectionModel_DB setMaxCandidates value
TextDetectionModel_DB setPolygonThreshold value
TextDetectionModel_DB setUnclipRatio value
TextDetectionModel_DB setInputParams scalefactor width height mean_color_list swapRB ?crop?
TextDetectionModel_DB close
TextDetectionModel_DB is the high-level API for text detection DL networks compatible with DB model. OpenCV 4.5.1 add the high-level API.
cv::dnn::TextRecognitionModel model ?config?
TextRecognitionModel recognize matrix
TextRecognitionModel getDecodeType
TextRecognitionModel getVocabulary
TextRecognitionModel setDecodeType value
TextRecognitionModel setVocabulary vocabulary
TextRecognitionModel setInputParams scalefactor width height mean_color_list swapRB ?crop?
TextRecognitionModel close
TextRecognitionModel is the high-level API for text recognition networks. OpenCV 4.5.1 add the high-level API.
Model files download links are provided in here.
::cv::thread::info ?tag?
::cv::thread::recv tag ?timeout?
::cv::thread::send tag mat|None ?string?
A simple mechanism to transfer cv::Mat from one Tcl thread to another
where tag is the name of a queue. The info command without arguments
returns all currently known tags, the info command with a tag returns
the number of items in that queue. With send an item consisting of a
cv::Mat or None and an optional string is added to the queue
named tag. With recv a zero, one, or two element list is returned
from the queue named tag with optional wait timeout in milliseconds
(default timeout is 10 seconds). Zero elements means timeout condition,
one or two elements are the name/command of the received cv::Mat or
empty element for None, and optional string.
This extension defined below namespace variables:
Types -
::cv::CV_8U
::cv::CV_8S
::cv::CV_16U
::cv::CV_16S
::cv::CV_32S
::cv::CV_32F
::cv::CV_64F
::cv::CV_8UC1
::cv::CV_8UC2
::cv::CV_8UC3
::cv::CV_8UC4
::cv::CV_8SC1
::cv::CV_8SC2
::cv::CV_8SC3
::cv::CV_8SC4
::cv::CV_16UC1
::cv::CV_16UC2
::cv::CV_16UC3
::cv::CV_16UC4
::cv::CV_16SC1
::cv::CV_16SC2
::cv::CV_16SC3
::cv::CV_16SC4
::cv::CV_32SC1
::cv::CV_32SC2
::cv::CV_32SC3
::cv::CV_32SC4
::cv::CV_32FC1
::cv::CV_32FC2
::cv::CV_32FC3
::cv::CV_32FC4
::cv::CV_64FC1
::cv::CV_64FC2
::cv::CV_64FC3
::cv::CV_64FC4
::cv::FileStorage flags -
::cv::FileStorage::READ
::cv::FileStorage::WRITE
::cv::FileStorage::APPEND
::cv::FileStorage::MEMORY
::cv::FileStorage::FORMAT_AUTO
::cv::FileStorage::FORMAT_XML
::cv::FileStorage::FORMAT_YAML
::cv::FileStorage::FORMAT_JSON
::cv::FileStorage::BASE64
::cv::FileStorage::WRITE_BASE64
::cv::sortIdx flags -
::cv::SORT_EVERY_ROW
::cv::SORT_EVERY_COLUMN
::cv::SORT_ASCENDING
::cv::SORT_DESCENDING
::cv::imread flags -
::cv::IMREAD_UNCHANGED
::cv::IMREAD_GRAYSCALE
::cv::IMREAD_COLOR
::cv::IMREAD_ANYDEPTH
::cv::IMREAD_ANYCOLOR
::cv::IMREAD_LOAD_GDAL
::cv::IMREAD_REDUCED_GRAYSCALE_2
::cv::IMREAD_REDUCED_COLOR_2
::cv::IMREAD_REDUCED_GRAYSCALE_4
::cv::IMREAD_REDUCED_COLOR_4
::cv::IMREAD_REDUCED_GRAYSCALE_8
::cv::IMREAD_REDUCED_COLOR_8
::cv::IMREAD_IGNORE_ORIENTATION
VideoCapture command flags -
::cv::CAP_ANY
::cv::CAP_VFW
::cv::CAP_V4L
::cv::CAP_V4L2
::cv::CAP_FIREWIRE
::cv::CAP_FIREWARE
::cv::CAP_IEEE1394
::cv::CAP_DC1394
::cv::CAP_CMU1394
::cv::CAP_QT
::cv::CAP_UNICAP
::cv::CAP_DSHOW
::cv::CAP_PVAPI
::cv::CAP_OPENNI
::cv::CAP_OPENNI_ASUS
::cv::CAP_ANDROID
::cv::CAP_XIAPI
::cv::CAP_AVFOUNDATION
::cv::CAP_GIGANETIX
::cv::CAP_MSMF
::cv::CAP_WINRT
::cv::CAP_INTELPERC
::cv::CAP_REALSENSE
::cv::CAP_OPENNI2
::cv::CAP_OPENNI2_ASUS
::cv::CAP_OPENNI2_ASTRA
::cv::CAP_GPHOTO2
::cv::CAP_GSTREAMER
::cv::CAP_FFMPEG
::cv::CAP_IMAGES
::cv::CAP_ARAVIS
::cv::CAP_OPENCV_MJPEG
::cv::CAP_INTEL_MFX
::cv::CAP_XINE
::cv::CAP_UEYE
VideoCapture and VideoWriter propId -
::cv::CAP_PROP_POS_MSEC
::cv::CAP_PROP_POS_FRAMES
::cv::CAP_PROP_POS_AVI_RATIO
::cv::CAP_PROP_FRAME_WIDTH
::cv::CAP_PROP_FRAME_HEIGHT
::cv::CAP_PROP_FPS
::cv::CAP_PROP_FOURCC
::cv::CAP_PROP_FRAME_COUNT
::cv::CAP_PROP_FORMAT
::cv::CAP_PROP_MODE
::cv::CAP_PROP_BRIGHTNESS
::cv::CAP_PROP_CONTRAST
::cv::CAP_PROP_SATURATION
::cv::CAP_PROP_HUE
::cv::CAP_PROP_GAIN
::cv::CAP_PROP_EXPOSURE
::cv::CAP_PROP_CONVERT_RGB
::cv::CAP_PROP_WHITE_BALANCE_BLUE_U
::cv::CAP_PROP_RECTIFICATION
::cv::CAP_PROP_MONOCHROME
::cv::CAP_PROP_SHARPNESS
::cv::CAP_PROP_AUTO_EXPOSURE
::cv::CAP_PROP_GAMMA
::cv::CAP_PROP_TEMPERATURE
::cv::CAP_PROP_TRIGGER
::cv::CAP_PROP_TRIGGER_DELAY
::cv::CAP_PROP_WHITE_BALANCE_RED_V
::cv::CAP_PROP_ZOOM
::cv::CAP_PROP_FOCUS
::cv::CAP_PROP_GUID
::cv::CAP_PROP_ISO_SPEED
::cv::CAP_PROP_BACKLIGHT
::cv::CAP_PROP_PAN
::cv::CAP_PROP_TILT
::cv::CAP_PROP_ROLL
::cv::CAP_PROP_IRIS
::cv::CAP_PROP_SETTINGS
::cv::CAP_PROP_BUFFERSIZE
::cv::CAP_PROP_AUTOFOCUS
::cv::CAP_PROP_SAR_NUM
::cv::CAP_PROP_SAR_DEN
::cv::CAP_PROP_BACKEND
::cv::CAP_PROP_CHANNEL
::cv::CAP_PROP_AUTO_WB
::cv::CAP_PROP_WB_TEMPERATURE
::cv::CAP_PROP_CODEC_PIXEL_FORMAT
::cv::CAP_PROP_BITRATE
::cv::CAP_PROP_ORIENTATION_META
::cv::CAP_PROP_ORIENTATION_AUTO
::cv::CAP_PROP_HW_ACCELERATION
::cv::CAP_PROP_HW_DEVICE
VideoWriter propId -
::cv::VIDEOWRITER_PROP_QUALITY
::cv::VIDEOWRITER_PROP_FRAMEBYTES
::cv::VIDEOWRITER_PROP_NSTRIPES
::cv::VIDEOWRITER_PROP_IS_COLOR
::cv::VIDEOWRITER_PROP_DEPTH
::cv::VIDEOWRITER_PROP_HW_ACCELERATION
::cv::VIDEOWRITER_PROP_HW_DEVICE
::cv::cvtColor code -
::cv::COLOR_BGR2BGRA
::cv::COLOR_RGB2RGBA
::cv::COLOR_BGRA2BGR
::cv::COLOR_RGBA2RGB
::cv::COLOR_BGR2RGBA
::cv::COLOR_RGB2BGRA
::cv::COLOR_RGBA2BGR
::cv::COLOR_BGRA2RGB
::cv::COLOR_BGR2RGB
::cv::COLOR_RGB2BGR
::cv::COLOR_BGRA2RGBA
::cv::COLOR_RGBA2BGRA
::cv::COLOR_BGR2GRAY
::cv::COLOR_RGB2GRAY
::cv::COLOR_GRAY2BGR
::cv::COLOR_GRAY2RGB
::cv::COLOR_GRAY2BGRA
::cv::COLOR_GRAY2RGBA
::cv::COLOR_BGRA2GRAY
::cv::COLOR_RGBA2GRAY
::cv::COLOR_BGR2BGR565
::cv::COLOR_RGB2BGR565
::cv::COLOR_BGR5652BGR
::cv::COLOR_BGR5652RGB
::cv::COLOR_BGRA2BGR565
::cv::COLOR_RGBA2BGR565
::cv::COLOR_BGR5652BGRA
::cv::COLOR_BGR5652RGBA
::cv::COLOR_GRAY2BGR565
::cv::COLOR_BGR5652GRAY
::cv::COLOR_BGR2BGR555
::cv::COLOR_RGB2BGR555
::cv::COLOR_BGR5552BGR
::cv::COLOR_BGR5552RGB
::cv::COLOR_BGRA2BGR555
::cv::COLOR_RGBA2BGR555
::cv::COLOR_BGR5552BGRA
::cv::COLOR_BGR5552RGBA
::cv::COLOR_GRAY2BGR555
::cv::COLOR_BGR5552GRAY
::cv::COLOR_BGR2XYZ
::cv::COLOR_RGB2XYZ
::cv::COLOR_XYZ2BGR
::cv::COLOR_XYZ2RGB
::cv::COLOR_BGR2YCrCb
::cv::COLOR_RGB2YCrCb
::cv::COLOR_YCrCb2BGR
::cv::COLOR_YCrCb2RGB
::cv::COLOR_BGR2HSV
::cv::COLOR_RGB2HSV
::cv::COLOR_BGR2Lab
::cv::COLOR_RGB2Lab
::cv::COLOR_BGR2Luv
::cv::COLOR_RGB2Luv
::cv::COLOR_BGR2HLS
::cv::COLOR_RGB2HLS
::cv::COLOR_HSV2BGR
::cv::COLOR_HSV2RGB
::cv::COLOR_Lab2BGR
::cv::COLOR_Lab2RGB
::cv::COLOR_Luv2BGR
::cv::COLOR_Luv2RGB
::cv::COLOR_HLS2BGR
::cv::COLOR_HLS2RGB
::cv::COLOR_BGR2HSV_FULL
::cv::COLOR_RGB2HSV_FULL
::cv::COLOR_BGR2HLS_FULL
::cv::COLOR_RGB2HLS_FULL
::cv::COLOR_HSV2BGR_FULL
::cv::COLOR_HSV2RGB_FULL
::cv::COLOR_HLS2BGR_FULL
::cv::COLOR_HLS2RGB_FULL
::cv::COLOR_LBGR2Lab
::cv::COLOR_LRGB2Lab
::cv::COLOR_LBGR2Luv
::cv::COLOR_LRGB2Luv
::cv::COLOR_Lab2LBGR
::cv::COLOR_Lab2LRGB
::cv::COLOR_Luv2LBGR
::cv::COLOR_Luv2LRGB
::cv::COLOR_BGR2YUV
::cv::COLOR_RGB2YUV
::cv::COLOR_YUV2BGR
::cv::COLOR_YUV2RGB
::cv::COLOR_YUV2RGB_NV12
::cv::COLOR_YUV2BGR_NV12
::cv::COLOR_YUV2RGB_NV21
::cv::COLOR_YUV2BGR_NV21
::cv::COLOR_YUV420sp2RGB
::cv::COLOR_YUV420sp2BGR
::cv::COLOR_YUV2RGBA_NV12
::cv::COLOR_YUV2BGRA_NV12
::cv::COLOR_YUV2RGBA_NV21
::cv::COLOR_YUV2BGRA_NV21
::cv::COLOR_YUV420sp2RGBA
::cv::COLOR_YUV420sp2BGRA
::cv::COLOR_YUV2RGB_YV12
::cv::COLOR_YUV2BGR_YV12
::cv::COLOR_YUV2RGB_IYUV
::cv::COLOR_YUV2BGR_IYUV
::cv::COLOR_YUV2RGB_I420
::cv::COLOR_YUV2BGR_I420
::cv::COLOR_YUV420p2RGB
::cv::COLOR_YUV420p2BGR
::cv::COLOR_YUV2RGBA_YV12
::cv::COLOR_YUV2BGRA_YV12
::cv::COLOR_YUV2RGBA_IYUV
::cv::COLOR_YUV2BGRA_IYUV
::cv::COLOR_YUV2RGBA_I420
::cv::COLOR_YUV2BGRA_I420
::cv::COLOR_YUV420p2RGBA
::cv::COLOR_YUV420p2BGRA
::cv::COLOR_YUV2GRAY_420
::cv::COLOR_YUV2GRAY_NV21
::cv::COLOR_YUV2GRAY_NV12
::cv::COLOR_YUV2GRAY_YV12
::cv::COLOR_YUV2GRAY_IYUV
::cv::COLOR_YUV2GRAY_I420
::cv::COLOR_YUV420sp2GRAY
::cv::COLOR_YUV420p2GRAY
::cv::COLOR_YUV2RGB_UYVY
::cv::COLOR_YUV2BGR_UYVY
::cv::COLOR_YUV2RGB_Y422
::cv::COLOR_YUV2BGR_Y422
::cv::COLOR_YUV2RGB_UYNV
::cv::COLOR_YUV2BGR_UYNV
::cv::COLOR_YUV2RGBA_UYVY
::cv::COLOR_YUV2BGRA_UYVY
::cv::COLOR_YUV2RGBA_Y422
::cv::COLOR_YUV2BGRA_Y422
::cv::COLOR_YUV2RGBA_UYNV
::cv::COLOR_YUV2BGRA_UYNV
::cv::COLOR_YUV2RGB_YUY2
::cv::COLOR_YUV2BGR_YUY2
::cv::COLOR_YUV2RGB_YVYU
::cv::COLOR_YUV2BGR_YVYU
::cv::COLOR_YUV2RGB_YUYV
::cv::COLOR_YUV2BGR_YUYV
::cv::COLOR_YUV2RGB_YUNV
::cv::COLOR_YUV2BGR_YUNV
::cv::COLOR_YUV2RGBA_YUY2
::cv::COLOR_YUV2BGRA_YUY2
::cv::COLOR_YUV2RGBA_YVYU
::cv::COLOR_YUV2BGRA_YVYU
::cv::COLOR_YUV2RGBA_YUYV
::cv::COLOR_YUV2BGRA_YUYV
::cv::COLOR_YUV2RGBA_YUNV
::cv::COLOR_YUV2BGRA_YUNV
::cv::COLOR_YUV2GRAY_UYVY
::cv::COLOR_YUV2GRAY_YUY2
::cv::COLOR_YUV2GRAY_Y422
::cv::COLOR_YUV2GRAY_UYNV
::cv::COLOR_YUV2GRAY_YVYU
::cv::COLOR_YUV2GRAY_YUYV
::cv::COLOR_YUV2GRAY_YUNV
::cv::COLOR_RGBA2mRGBA
::cv::COLOR_mRGBA2RGBA
::cv::COLOR_RGB2YUV_I420
::cv::COLOR_BGR2YUV_I420
::cv::COLOR_RGB2YUV_IYUV
::cv::COLOR_BGR2YUV_IYUV
::cv::COLOR_RGBA2YUV_I420
::cv::COLOR_BGRA2YUV_I420
::cv::COLOR_RGBA2YUV_IYUV
::cv::COLOR_BGRA2YUV_IYUV
::cv::COLOR_RGB2YUV_YV12
::cv::COLOR_BGR2YUV_YV12
::cv::COLOR_RGBA2YUV_YV12
::cv::COLOR_BGRA2YUV_YV12
::cv::COLOR_BayerBG2BGR
::cv::COLOR_BayerGB2BGR
::cv::COLOR_BayerRG2BGR
::cv::COLOR_BayerGR2BGR
::cv::COLOR_BayerBG2RGB
::cv::COLOR_BayerGB2RGB
::cv::COLOR_BayerRG2RGB
::cv::COLOR_BayerGR2RGB
::cv::COLOR_BayerBG2GRAY
::cv::COLOR_BayerGB2GRAY
::cv::COLOR_BayerRG2GRAY
::cv::COLOR_BayerGR2GRAY
::cv::COLOR_BayerBG2BGR_VNG
::cv::COLOR_BayerGB2BGR_VNG
::cv::COLOR_BayerRG2BGR_VNG
::cv::COLOR_BayerGR2BGR_VNG
::cv::COLOR_BayerBG2RGB_VNG
::cv::COLOR_BayerGB2RGB_VNG
::cv::COLOR_BayerRG2RGB_VNG
::cv::COLOR_BayerGR2RGB_VNG
::cv::COLOR_BayerBG2BGR_EA
::cv::COLOR_BayerGB2BGR_EA
::cv::COLOR_BayerRG2BGR_EA
::cv::COLOR_BayerGR2BGR_EA
::cv::COLOR_BayerBG2RGB_EA
::cv::COLOR_BayerGB2RGB_EA
::cv::COLOR_BayerRG2RGB_EA
::cv::COLOR_BayerGR2RGB_EA
::cv::COLOR_BayerBG2BGRA
::cv::COLOR_BayerGB2BGRA
::cv::COLOR_BayerRG2BGRA
::cv::COLOR_BayerGR2BGRA
::cv::COLOR_BayerBG2RGBA
::cv::COLOR_BayerGB2RGBA
::cv::COLOR_BayerRG2RGBA
::cv::COLOR_BayerGR2RGBA
::cv::COLOR_COLORCVT_MAX
Line Types -
::cv::FILLED
::cv::LINE_4
::cv::LINE_8
::cv::LINE_AA
::cv::drawMarker Marker Types -
::cv::MARKER_CROSS
::cv::MARKER_TILTED_CROSS
::cv::MARKER_STAR
::cv::MARKER_DIAMOND
::cv::MARKER_SQUARE
::cv::MARKER_TRIANGLE_UP
::cv::MARKER_TRIANGLE_DOWN
::cv::putText fontFace -
::cv::FONT_HERSHEY_SIMPLEX
::cv::FONT_HERSHEY_PLAIN
::cv::FONT_HERSHEY_DUPLEX
::cv::FONT_HERSHEY_COMPLEX
::cv::FONT_HERSHEY_TRIPLEX
::cv::FONT_HERSHEY_COMPLEX_SMALL
::cv::FONT_HERSHEY_SCRIPT_SIMPLEX
::cv::FONT_HERSHEY_SCRIPT_COMPLEX
::cv::FONT_ITALIC
::cv::threshold type -
::cv::THRESH_BINARY
::cv::THRESH_BINARY_INV
::cv::THRESH_TRUNC
::cv::THRESH_TOZERO
::cv::THRESH_TOZERO_INV
::cv::THRESH_MASK
::cv::THRESH_OTSU
::cv::THRESH_TRIANGLE
::cv::adaptiveThreshold method -
::cv::ADAPTIVE_THRESH_MEAN_C
::cv::ADAPTIVE_THRESH_GAUSSIAN_C
::cv::copyMakeBorder border type -
::cv::BORDER_CONSTANT
::cv::BORDER_REPLICATE
::cv::BORDER_REFLECT
::cv::BORDER_WRAP
::cv::BORDER_REFLECT_101
::cv::BORDER_TRANSPARENT
::cv::BORDER_REFLECT101
::cv::BORDER_DEFAULT
::cv::BORDER_ISOLATED
::cv::getStructuringElement morph shapes -
::cv::MORPH_RECT
::cv::MORPH_CROSS
::cv::MORPH_ELLIPSE
::cv::morphologyEx op type -
::cv::MORPH_ERODE
::cv::MORPH_DILATE
::cv::MORPH_OPEN
::cv::MORPH_CLOSE
::cv::MORPH_GRADIENT
::cv::MORPH_TOPHAT
::cv::MORPH_BLACKHAT
::cv::MORPH_HITMISS
::cv::matchShapes method -
::cv::CONTOURS_MATCH_I1
::cv::CONTOURS_MATCH_I2
::cv::CONTOURS_MATCH_I3
::cv::applyColorMap colormap -
::cv::COLORMAP_AUTUMN
::cv::COLORMAP_BONE
::cv::COLORMAP_JET
::cv::COLORMAP_WINTER
::cv::COLORMAP_RAINBOW
::cv::COLORMAP_OCEAN
::cv::COLORMAP_SUMMER
::cv::COLORMAP_SPRING
::cv::COLORMAP_COOL
::cv::COLORMAP_HSV
::cv::COLORMAP_PINK
::cv::COLORMAP_HOT
::cv::COLORMAP_PARULA
::cv::COLORMAP_MAGMA
::cv::COLORMAP_INFERNO
::cv::COLORMAP_PLASMA
::cv::COLORMAP_VIRIDIS
::cv::COLORMAP_CIVIDIS
::cv::COLORMAP_TWILIGHT
::cv::COLORMAP_TWILIGHT_SHIFTED
::cv::COLORMAP_TURBO
::cv::COLORMAP_DEEPGREEN
::cv::namedWindow flags -
::cv::WINDOW_NORMAL
::cv::WINDOW_AUTOSIZE
::cv::WINDOW_OPENGL
::cv::WINDOW_FULLSCREEN
::cv::WINDOW_FREERATIO
::cv::WINDOW_KEEPRATIO
::cv::WINDOW_GUI_EXPANDED
::cv::WINDOW_GUI_NORMAL
::cv::setMouseCallback event -
::cv::EVENT_MOUSEMOVE
::cv::EVENT_LBUTTONDOWN
::cv::EVENT_RBUTTONDOWN
::cv::EVENT_MBUTTONDOWN
::cv::EVENT_LBUTTONUP
::cv::EVENT_RBUTTONUP
::cv::EVENT_MBUTTONUP
::cv::EVENT_LBUTTONDBLCLK
::cv::EVENT_RBUTTONDBLCLK
::cv::EVENT_MBUTTONDBLCLK
::cv::EVENT_MOUSEWHEEL
::cv::EVENT_MOUSEHWHEEL
::cv::EVENT_FLAG_LBUTTON
::cv::EVENT_FLAG_RBUTTON
::cv::EVENT_FLAG_MBUTTON
::cv::EVENT_FLAG_CTRLKEY
::cv::EVENT_FLAG_SHIFTKEY
::cv::EVENT_FLAG_ALTKEY
::cv::compareHist method -
::cv::HISTCMP_CORREL
::cv::HISTCMP_CHISQR
::cv::HISTCMP_INTERSECT
::cv::HISTCMP_BHATTACHARYYA
::cv::HISTCMP_HELLINGER
::cv::HISTCMP_CHISQR_ALT
::cv::HISTCMP_KL_DIV
LineSegmentDetector modes -
::cv::LSD_REFINE_NONE
::cv::LSD_REFINE_STD
::cv::LSD_REFINE_ADV
::cv::dft command flags -
::cv::DFT_INVERSE
::cv::DFT_SCALE
::cv::DFT_ROWS
::cv::DFT_COMPLEX_OUTPUT
::cv::DFT_REAL_OUTPUT
::cv::DFT_COMPLEX_INPUT
::cv::DCT_INVERSE
::cv::DCT_ROWS
::cv::SVDecomp flags -
::cv::SVD_MODIFY_A
::cv::SVD_NO_UV
::cv::SVD_FULL_UV
::cv::norm, ::cv::normalize and BFMatchernorm_type -
::cv::NORM_INF
::cv::NORM_L1
::cv::NORM_L2
::cv::NORM_L2SQR
::cv::NORM_HAMMING
::cv::NORM_HAMMING2
::cv::NORM_TYPE_MASK
::cv::NORM_RELATIVE
::cv::NORM_MINMAX
::cv::floodFill command flag -
::cv::FLOODFILL_FIXED_RANGE
:cv::remapinterpolation,
::cv::resize, ::cv::warpAffine and ::cv::warpPerspective command flag -
::cv::INTER_NEAREST
::cv::INTER_LINEAR
::cv::INTER_CUBIC
::cv::INTER_AREA
::cv::INTER_LANCZOS4
::cv::INTER_LINEAR_EXACT
::cv::INTER_NEAREST_EXACT
::cv::INTER_MAX
::cv::WARP_FILL_OUTLIERS
::cv::WARP_INVERSE_MAP
::cv::warpPolar mode -
::cv::WARP_POLAR_LINEAR
::cv::WARP_POLAR_LOG
::cv::compare cmpop -
::cv::CMP_EQ
::cv::CMP_GT
::cv::CMP_GE
::cv::CMP_LT
::cv::CMP_LE
::cv::CMP_NE
::cv::reduce rtype -
::cv::REDUCE_SUM
::cv::REDUCE_AVG
::cv::REDUCE_MAX
::cv::REDUCE_MIN
::cv::REDUCE_SUM2
OpenCV 4.8.0 add ::cv::REDUCE_SUM2 option to ::cv::reduce.
::cv::rotate rotateCode -
::cv::ROTATE_90_CLOCKWISE
::cv::ROTATE_180
::cv::ROTATE_90_COUNTERCLOCKWISE
MATRIX inv method, ::cv::solve flags and
::cv::getPerspectiveTransform command solveMethod -
::cv::DECOMP_LU
::cv::DECOMP_SVD
::cv::DECOMP_EIG
::cv::DECOMP_CHOLESKY
::cv::DECOMP_QR
::cv::DECOMP_NORMAL
::cv::findContours mode -
::cv::RETR_EXTERNAL
::cv::RETR_LIST
::cv::RETR_CCOMP
::cv::RETR_TREE
::cv::RETR_FLOODFILL
::cv::findContours method -
::cv::CHAIN_APPROX_NONE
::cv::CHAIN_APPROX_SIMPLE
::cv::CHAIN_APPROX_TC89_L1
::cv::CHAIN_APPROX_TC89_KCOS
::cv::HoughCircles method -
::cv::HOUGH_GRADIENT
::cv::HOUGH_GRADIENT_ALT
::cv::matchTemplate method -
::cv::TM_SQDIFF
::cv::TM_SQDIFF_NORMED
::cv::TM_CCORR
::cv::TM_CCORR_NORMED
::cv::TM_CCOEFF
::cv::TM_CCOEFF_NORMED
::cv::EMD distType, ::cv::fitLine distType and
::cv::distanceTransform distanceType -
::cv::DIST_L1
::cv::DIST_L2
::cv::DIST_C
::cv::DIST_L12
::cv::DIST_FAIR
::cv::DIST_WELSCH
::cv::DIST_HUBER
GrabCut classes -
::cv::GC_BGD
::cv::GC_FGD
::cv::GC_PR_BGD
::cv::GC_PR_FGD
::cv::calcCovarMatrix flags -
::cv::COVAR_SCRAMBLED
::cv::COVAR_NORMAL
::cv::COVAR_USE_AVG
::cv::COVAR_SCALE
::cv::COVAR_ROWS
::cv::COVAR_COLS
Kmeans flags -
::cv::KMEANS_RANDOM_CENTERS
::cv::KMEANS_PP_CENTERS
::cv::KMEANS_USE_INITIAL_LABELS
PCA flags -
::cv::DATA_AS_ROW
::cv::DATA_AS_COL
::cv::USE_AVG
TermCriteria type -
::cv::COUNT
::cv::EPS
FastFeatureDetector detector type -
::cv::DetectorType_TYPE_5_8
::cv::DetectorType_TYPE_7_12
::cv::DetectorType_TYPE_9_16
AgastFeatureDetector detector type -
::cv::DetectorType_AGAST_5_8
::cv::DetectorType_AGAST_7_12d
::cv::DetectorType_AGAST_7_12s
::cv::DetectorType_OAST_9_16
ORB Score type -
::cv::ORB_HARRIS_SCORE
::cv::ORB_FAST_SCORE
AKAZE descriptor type -
::cv::AKAZE_DESCRIPTOR_KAZE_UPRIGHT
::cv::AKAZE_DESCRIPTOR_KAZE
::cv::AKAZE_DESCRIPTOR_MLDB_UPRIGHT
::cv::AKAZE_DESCRIPTOR_MLDB
KAZE and AKAZE Diffusivity type -
::cv::KAZE_DIFF_PM_G1
::cv::KAZE_DIFF_PM_G2
::cv::KAZE_DIFF_WEICKERT
::cv::KAZE_DIFF_CHARBONNIER
drawKeypoints and drawMatches flags
::cv::DRAW_MATCHES_FLAGS_DEFAULT
::cv::DRAW_MATCHES_FLAGS_DRAW_OVER_OUTIMG
::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS
::cv::DRAW_MATCHES_FLAGS_DRAW_RICH_KEYPOINTS
::cv::findTransformECC motion type -
::cv::MOTION_TRANSLATION
::cv::MOTION_EUCLIDEAN
::cv::MOTION_AFFINE
::cv::MOTION_HOMOGRAPHY
findChessboardCorners flags -
::cv::CALIB_CB_ADAPTIVE_THRESH
::cv::CALIB_CB_NORMALIZE_IMAGE
::cv::CALIB_CB_FILTER_QUADS
::cv::CALIB_CB_FAST_CHECK
findFundamentalMat method -
::cv::FM_7POINT
::cv::FM_8POINT
::cv::FM_LMEDS
::cv::FM_RANSAC
findHomography method -
::cv::LMEDS
::cv::RANSAC
::cv::RHO
StereoBM PreFilter type -
::cv::PREFILTER_NORMALIZED_RESPONSE
::cv::PREFILTER_XSOBEL
StereoSGBM mode -
::cv::StereoSGBM_MODE_SGBM
::cv::StereoSGBM_MODE_HH
::cv::StereoSGBM_MODE_SGBM_3WAY
::cv::StereoSGBM_MODE_HH4
QRCodeEncoder CorrectionLevel mode -
::cv::QRCodeEncoder_CORRECT_LEVEL_L
::cv::QRCodeEncoder_CORRECT_LEVEL_M
::cv::QRCodeEncoder_CORRECT_LEVEL_Q
::cv::QRCodeEncoder_CORRECT_LEVEL_H
QRCodeEncoder EncodeMode mode -
::cv::QRCodeEncoder_MODE_AUTO
::cv::QRCodeEncoder_MODE_NUMERIC
::cv::QRCodeEncoder_MODE_ALPHANUMERIC
::cv::QRCodeEncoder_MODE_BYTE
::cv::QRCodeEncoder_MODE_ECI
::cv::QRCodeEncoder_MODE_KANJI
::cv::QRCodeEncoder_MODE_STRUCTURED_APPEND
Photo inpaint flags -
::cv::INPAINT_NS
::cv::INPAINT_TELEA
Photo edgePreservingFilter flags -
::cv::RECURS_FILTER
::cv::NORMCONV_FILTER
Photo seamlessClone flags -
::cv::NORMAL_CLONE
::cv::MIXED_CLONE
::cv::MONOCHROME_TRANSFER
Stitcher mode -
::cv::PANORAMA
::cv::SCANS
::cv::FaceRecognizerSF dis_type -
::cv::FR_COSINE
::cv::FR_NORM_L2
ML sample types -
::cv::ml::ROW_SAMPLE
::cv::ml::COL_SAMPLE
Predict options -
::cv::ml::UPDATE_MODEL
::cv::ml::RAW_OUTPUT
::cv::ml::COMPRESSED_INPUT
::cv::ml::PREPROCESSED_INPUT
LogisticRegression Methods -
::cv::ml::LOGISTIC_BATCH
::cv::ml::LOGISTIC_MINI_BATCH
LogisticRegression Regularization Kinds -
::cv::ml::LOGISTIC_REG_DISABLE
::cv::ml::LOGISTIC_REG_L1
::cv::ml::LOGISTIC_REG_L2
KNearest algorithm -
::cv::ml::KNEAREST_BRUTE_FORCE
::cv::ml::KNEAREST_KDTREE
SVM types -
::cv::ml::SVM_C_SVC
::cv::ml::SVM_NU_SVC
::cv::ml::SVM_ONE_CLASS
::cv::ml::SVM_EPS_SVR
::cv::ml::SVM_NU_SVR
SVM kernel types -
::cv::ml::SVM_LINEAR
::cv::ml::SVM_POLY
::cv::ml::SVM_RBF
::cv::ml::SVM_SIGMOID
::cv::ml::SVM_CHI2
::cv::ml::SVM_INTER
SVMSGD Margin Type -
::cv::ml::SVMSGD_SOFT_MARGIN
::cv::ml::SVMSGD_HARD_MARGIN
SVMSGD Type -
::cv::ml::SVMSGD_SGD
::cv::ml::SVMSGD_ASGD
Boost Types -
::cv::ml::BOOST_DISCRETE
::cv::ml::BOOST_REAL
::cv::ml::BOOST_LOGIT
::cv::ml::BOOST_GENTLE
ANN_MLP ActivationFunctions -
::cv::ml::MLP_IDENTITY
::cv::ml::MLP_SIGMOID_SYM
::cv::ml::MLP_GAUSSIAN
::cv::ml::MLP_RELU
::cv::ml::MLP_LEAKYRELU
ANN_MLP Train Flags
::cv::ml::MLP_UPDATE_WEIGHTS
::cv::ml::MLP_NO_INPUT_SCALE
::cv::ml::MLP_NO_OUTPUT_SCALE
ANN_MLP Training Methods -
::cv::ml::MLP_BACKPROP
::cv::ml::MLP_RPROP
::cv::ml::MLP_ANNEAL
READNET setPreferableBackend backendId -
::cv::dnn::DNN_BACKEND_DEFAULT
::cv::dnn::DNN_BACKEND_HALIDE
::cv::dnn::DNN_BACKEND_OPENCV
::cv::dnn::DNN_BACKEND_VKCOM
::cv::dnn::DNN_BACKEND_CUDA
READNET setPreferableTarget targetId -
::cv::dnn::DNN_TARGET_CPU
::cv::dnn::DNN_TARGET_OPENCL
::cv::dnn::DNN_TARGET_OPENCL_FP16
::cv::dnn::DNN_TARGET_MYRIAD
::cv::dnn::DNN_TARGET_VULKAN
::cv::dnn::DNN_TARGET_CUDA
::cv::dnn::DNN_TARGET_CUDA_FP16
::cv::dnn::DNN_TARGET_HDDL
cv::dnn::softNMSBoxes method -
cv::dnn::SOFTNMS_LINEAR
cv::dnn::SOFTNMS_GAUSSIAN
Brighness and contrast -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set img2 [$img convertTo -1 1 100]
set img3 [$img convertTo -1 1 -100]
set img4 [$img convertTo -1 2 0]
set img5 [$img convertTo -1 0.5 0]
::cv::namedWindow "Source" $::cv::WINDOW_NORMAL
::cv::imshow "Source" $img
::cv::namedWindow "Brighness High" $::cv::WINDOW_NORMAL
::cv::imshow "Brighness High" $img2
::cv::namedWindow "Brighness Low" $::cv::WINDOW_NORMAL
::cv::imshow "Brighness Low" $img3
::cv::namedWindow "Contrast High" $::cv::WINDOW_NORMAL
::cv::imshow "Contrast High" $img4
::cv::namedWindow "Contrast Low" $::cv::WINDOW_NORMAL
::cv::imshow "Contrast Low" $img5
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$img2 close
$img3 close
$img4 close
$img5 close
} on error {em} {
puts $em
}
Below is an example to apply a color map -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set img2 [::cv::applyColorMap $img $::cv::COLORMAP_RAINBOW]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img
::cv::namedWindow "Display Image 2" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image 2" $img2
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$img2 close
} on error {em} {
puts $em
}
Flip and Rotate -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
# 0 means flipping around the x-axis
set f_img [::cv::flip $img 0]
# Rotate by 270 degrees clockwise
set r_img [::cv::rotate $img $::cv::ROTATE_90_COUNTERCLOCKWISE]
::cv::namedWindow "Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Image" $img
::cv::namedWindow "Flip" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Flip" $f_img
::cv::namedWindow "Rotate" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Rotate" $r_img
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$f_img close
$r_img close
} on error {em} {
puts $em
}
Image Rotation and Translation -
package require opencv
if {$argc != 1} {
puts "Please give a file name."
}
set filename [lindex $argv 0]
try {
set img1 [cv::imread $filename]
# Rotation
set matrix1 [::cv::getRotationMatrix2D [expr [$img1 cols]/2] [expr [$img1 rows]/2] 45 1]
set img2 [::cv::warpAffine $img1 $matrix1 [$img1 cols] [$img1 rows]]
# Translation
set matrix2 [cv::Mat::Mat 2 3 $::cv::CV_64FC1]
$matrix2 setData [list 1.0 0.0 100 0.0 1.0 100]
set img3 [::cv::warpAffine $img1 $matrix2 [$img1 cols] [$img1 rows]]
# Wrap
set cols [$img1 cols]
set rows [$img1 rows]
set matrix3 [::cv::getAffineTransform \
[list 0 0 [expr $cols-1] 0 0 [expr $rows-1]] \
[list 0 [expr $rows * 0.33] [expr $cols*0.85] [expr $rows*0.25] \
[expr $cols*0.15] [expr $rows*0.7]]]
set img4 [::cv::warpAffine $img1 $matrix3 [$img1 cols] [$img1 rows]]
# Output
::cv::imwrite "rotation.jpg" $img2
::cv::imwrite "translation.jpg" $img3
::cv::imwrite "wrap.jpg" $img4
$matrix1 close
$matrix2 close
$matrix3 close
$img1 close
$img2 close
$img3 close
$img4 close
} on error {em} {
puts $em
}
Morphological Operations example -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set kernel [::cv::getStructuringElement $::cv::MORPH_ELLIPSE 5 5]
set dst [::cv::morphologyEx $img $::cv::MORPH_ERODE $kernel]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img
::cv::namedWindow "Display Image 2" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image 2" $dst
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$dst close
$kernel close
} on error {em} {
puts $em
}
Below is an exmaple for Laplacian Operator:
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
# Remove noise by applying a Gaussian blur
set img2 [::cv::GaussianBlur $img 3 3 0 0 $::cv::BORDER_DEFAULT]
# Convert the original image to grayscale
set img3 [::cv::cvtColor $img2 $::cv::COLOR_BGR2GRAY]
# Applies a Laplacian operator to the grayscale image
set img4 [::cv::Laplacian $img3 3 1.0 0.0 $::cv::BORDER_DEFAULT]
# Display
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::moveWindow "Display Image" 0 0
::cv::imshow "Display Image" $img4
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$img2 close
$img3 close
$img4 close
} on error {em} {
puts $em
}
Below is an example that apply OpenCV Look Up Table (LUT) to an image -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
set luv_mat [cv::Mat::Mat 1 256 $::cv::CV_8UC1]
for {set i 0} {$i < 256} {incr i} {
if {$i > 64 && $i < 196} {
$luv_mat at [list 0 $i] 0 0
} else {
$luv_mat at [list 0 $i] 0 $i
}
}
set img2 [::cv::LUT $img $luv_mat]
::cv::imshow "Display Image" $img2
::cv::waitKey 0
::cv::destroyAllWindows
$luv_mat close
$img close
$img2 close
} on error {em} {
puts $em
}
Below is an example to add 2 images:
package require opencv
if {$argc != 2} {
puts "Please give 2 image file name."
exit
}
set file1 [lindex $argv 0]
set file2 [lindex $argv 1]
try {
set img1 [::cv::imread $file1 $::cv::IMREAD_COLOR]
set img2 [::cv::imread $file2 $::cv::IMREAD_COLOR]
if {[$img1 rows] < [$img2 rows] || [$img1 cols] < [$img2 cols]} {
puts "Image 1 has to bigger than image 2."
}
set img1rect [$img1 rect 0 0 [$img2 cols] [$img2 rows]]
set addimage [::cv::addWeighted $img1rect 0.5 $img2 0.5 0]
$addimage copyTo $img1rect
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img1
::cv::waitKey 0
::cv::destroyAllWindows
$img1rect close
$addimage close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Play a video file -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
set v [::cv::VideoCapture file $filename]
if {[$v isOpened]==0} {
puts "Open Video file $filename failed."
exit
} else {
set fps [$v get $::cv::CAP_PROP_FPS]
puts "Frames per second : $fps"
set frame_count [$v get $::cv::CAP_PROP_FRAME_COUNT]
puts "Frame count : $frame_count"
}
while {[$v isOpened]==1} {
try {
set f [$v read]
::cv::imshow "Frame" $f
$f close
} on error {em} {
break
}
set key [::cv::waitKey 10]
if {$key==[scan "q" %c] || $key == 27} {
break
}
}
$v close
::cv::destroyAllWindows
Access camera using OpenCV and save to a video file -
package require opencv
if {$argc != 1} {
exit
}
set index [lindex $argv 0]
set v [::cv::VideoCapture index $index]
if {[$v isOpened]==0} {
puts "Open camera $index failed."
exit
}
set width [$v get 3]
set height [$v get 4]
set w [::cv::VideoWriter output.avi MJPG 20.0 $width $height 1]
while {[$v isOpened]==1} {
try {
set f [$v read]
::cv::imshow "Frame" $f
$w write $f
$f close
} on error {em} {
puts $em
break
}
set key [::cv::waitKey 1]
if {$key==[scan "q" %c]} {
break
}
}
$w close
$v close
::cv::destroyAllWindows
Video CamShift example -
package require opencv
# The example file can be downloaded from:
# https://www.bogotobogo.com/python/OpenCV_Python/images/mean_shift_tracking/slow_traffic_small.mp4
set filename "slow_traffic_small.mp4"
set v [::cv::VideoCapture file $filename]
if {[$v isOpened]==0} {
puts "Open Video file $filename failed."
exit
}
# Setup initial location of track window, simply hardcoded the values
set x 300
set y 200
set width 100
set height 50
# calcHist parameters
set channels [list 0]
set histSize [list 180]
set ranges [list 0 180]
set frame [$v read]
set roi [$frame crop $x $y $width $height]
set hsv_roi [::cv::cvtColor $roi $::cv::COLOR_BGR2HSV]
set mask [::cv::inRange $hsv_roi [list 0 60 32 0] [list 180 255 255 0]]
set hsv_hist [::cv::calcHist $hsv_roi $channels $mask 1 $histSize $ranges]
set roi_hist [::cv::normalize $hsv_hist 0 255 $::cv::NORM_MINMAX]
$frame close
$roi close
$hsv_roi close
$mask close
$hsv_hist close
set term [::cv::TermCriteria [expr $::cv::EPS | $::cv::COUNT] 10 1]
while {[$v isOpened]==1} {
try {
set frame [$v read]
if {$x < 550 || $y < 200} {
set hsv [cv::cvtColor $frame $::cv::COLOR_BGR2HSV]
set dst [cv::calcBackProject $hsv [list 0] $roi_hist [list 0 180]]
set result [::cv::CamShift $dst $x $y $width $height $term]
set x [lindex [lindex $result 0] 0]
set y [lindex [lindex $result 0] 1]
set width [lindex [lindex $result 0] 2]
set hehgit [lindex [lindex $result 0] 3]
set points [lindex $result 1]
set p0_x [lindex $points 0]
set p0_y [lindex $points 1]
set p1_x [lindex $points 2]
set p1_y [lindex $points 3]
set p2_x [lindex $points 4]
set p2_y [lindex $points 5]
set p3_x [lindex $points 6]
set p3_y [lindex $points 7]
::cv::line $frame $p0_x $p0_y $p1_x $p1_y [list 255 0 0 0] 3
::cv::line $frame $p1_x $p1_y $p2_x $p2_y [list 255 0 0 0] 3
::cv::line $frame $p2_x $p2_y $p3_x $p3_y [list 255 0 0 0] 3
::cv::line $frame $p3_x $p3_y $p0_x $p0_y [list 255 0 0 0] 3
$hsv close
$dst close
}
::cv::imshow "Frame" $frame
$frame close
} on error {em} {
break
}
set key [::cv::waitKey 10]
if {$key==[scan "q" %c] || $key == 27} {
break
}
}
$term close
$v close
::cv::destroyAllWindows
Lucas-Kanade Optical Flow in OpenCV -
package require opencv
# The example file can be downloaded from:
# https://www.bogotobogo.com/python/OpenCV_Python/images/mean_shift_tracking/slow_traffic_small.mp4
set filename "slow_traffic_small.mp4"
set v [::cv::VideoCapture file $filename]
if {[$v isOpened]==0} {
puts "Open Video file $filename failed."
exit
}
# Create some random colors
set colors [list]
for {set i 0} {$i < 100} {incr i} {
lappend colors [list [expr int(rand() * 255)] [expr int(rand() * 255)] [expr int(rand() * 255)] 0]
}
set frame [$v read]
set oldGray [::cv::cvtColor $frame $::cv::COLOR_BGR2GRAY]
set emptymask [::cv::Mat::Mat 0 0 $::cv::CV_8U]
set p0 [::cv::goodFeaturesToTrack $oldGray 100 0.3 7 $emptymask 7 0 0.04]
# Create a mask image for drawing purposes
set mask [::cv::Mat::Mat [$frame rows] [$frame cols] [$frame type] [list 0 0 0 0]]
$emptymask close
$frame close
while {[$v isOpened]==1} {
try {
set term [::cv::TermCriteria [expr $::cv::EPS | $::cv::COUNT] 10 0.03]
set frame [$v read]
set frameGray [::cv::cvtColor $frame $::cv::COLOR_BGR2GRAY]
set result [::cv::calcOpticalFlowPyrLK $oldGray $frameGray $p0 15 15 2 $term]
set p1 [lindex $result 0]
set status [lindex $result 1]
set err [lindex $result 2]
for {set i 0} {$i < [$p1 rows]} {incr i} {
for {set j 0} {$j < [$p1 cols]} {incr j} {
set oldx [expr int([$p0 at [list $i $j] 0])]
set oldy [expr int([$p0 at [list $i $j] 1])]
set x [expr int([$p1 at [list $i $j] 0])]
set y [expr int([$p1 at [list $i $j] 1])]
if {[$status at [list $i $j] 0]==1} {
::cv::line $mask $x $y $oldx $oldy [lindex $colors $i] 2
::cv::circle $frame $x $y 5 [lindex $colors $i] -1
}
}
}
set frame2 [::cv::add $frame $mask]
::cv::imshow "Frame" $frame2
$err close
$status close
$p0 close
$oldGray close
set p0 $p1
set oldGray $frameGray
$frame close
$frame2 close
$term close
} on error {em} {
break
}
set key [::cv::waitKey 10]
if {$key==[scan "q" %c] || $key == 27} {
break
}
}
$p1 close
$frameGray close
$v close
::cv::destroyAllWindows
Dense Optical Flow Example -
package require opencv
# The example file can be downloaded from:
# https://github.com/opencv/opencv/blob/master/samples/data/vtest.avi
set filename "vtest.avi"
set v [::cv::VideoCapture file $filename]
if {[$v isOpened]==0} {
puts "Open Video file $filename failed."
exit
}
set frame [$v read]
set oldGray [::cv::cvtColor $frame $::cv::COLOR_BGR2GRAY]
$frame close
while {[$v isOpened]==1} {
try {
set frame [$v read]
set frameGray [::cv::cvtColor $frame $::cv::COLOR_BGR2GRAY]
set flow [cv::calcOpticalFlowFarneback $oldGray $frameGray 0.5 3 15 3 5 1.2 0]
set flowlist [cv::split $flow]
set uvalue [lindex $flowlist 0]
set vvalue [lindex $flowlist 1]
set result [cv::cartToPolar $uvalue $vvalue 1]
set mag [lindex $result 0]
set ang [lindex $result 1]
set hsv0 [$ang multiply [expr ((1.0 / 360.0) * (180.0 / 255.0))]]
set hsv1 [::cv::Mat::ones [$frame rows] [$frame cols] $::cv::CV_32F]
set hsv2 [cv::normalize $mag 0 1 $::cv::NORM_MINMAX]
set hsv [::cv::merge [list $hsv0 $hsv1 $hsv2]]
set hsv8 [$hsv convertTo $::cv::CV_8U 255.0 0]
set frame2 [::cv::cvtColor $hsv8 $::cv::COLOR_HSV2BGR]
::cv::imshow "Frame" $frame2
$flow close
$uvalue close
$vvalue close
$mag close
$ang close
$hsv0 close
$hsv1 close
$hsv2 close
$hsv close
$hsv8 close
$oldGray close
set oldGray $frameGray
$frame close
$frame2 close
} on error {em} {
break
}
set key [::cv::waitKey 10]
if {$key==[scan "q" %c] || $key == 27} {
break
}
}
$frameGray close
$v close
::cv::destroyAllWindows
Below is a selectROI example -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set rect [::cv::selectROI $img]
set x [lindex $rect 0]
set y [lindex $rect 1]
set width [lindex $rect 2]
set height [lindex $rect 3]
# Create a mask for the ROI
set mask [::cv::Mat::zeros [$img rows] [$img cols] [$img type]]
::cv::rectangle $mask $x $y [expr $x+$width] [expr $y+$height] [list 255 255 255 0] -1
set img2 [::cv::colorChange $img $mask 2.0 0.5 0.5]
::cv::namedWindow "Display Result" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Result" $img2
::cv::waitKey 0
::cv::destroyAllWindows
$img close
$mask close
$img2 close
} on error {em} {
puts $em
}
Below is a setMouseCallback test -
package require opencv
if {$argc != 1} {
exit
}
proc drawit {event x y flags} {
if {$event == $::cv::EVENT_LBUTTONUP} {
set scalar [list 0 0 255 0]
::cv::circle $::img $x $y 25 $scalar 3 $::cv::LINE_8 0
::cv::imshow "Display Image" $::img
}
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::setMouseCallback "Display Image" drawit
::cv::imshow "Display Image" $img
::cv::waitKey 0
::cv::destroyAllWindows
$img close
} on error {em} {
puts $em
}
Below is a createTrackbar test -
package require opencv
if {$argc != 1} {
exit
}
proc changeit {value} {
set img2 [::cv::Sobel $::img 1 0 $value 1.0 0.0 $::cv::BORDER_DEFAULT]
::cv::imshow "Display Image" $img2
$img2 close
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::createTrackbar "Trackbar1" "Display Image" 3 21 changeit
changeit 3
::cv::waitKey 0
::cv::destroyAllWindows
$img close
} on error {em} {
puts $em
}
Contours test -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set image1 [::cv::imread $filename]
set image2 [::cv::cvtColor $image1 $::cv::COLOR_RGB2GRAY]
set image3 [::cv::threshold $image2 20 255 $::cv::THRESH_BINARY]
set contours [::cv::findContours $image3 $::cv::RETR_EXTERNAL $::cv::CHAIN_APPROX_SIMPLE]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $image1
::cv::namedWindow "Display Image 2" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image 2" $image3
set length [llength $contours]
set max 0
set maxindex 0
set bounding_rect [list]
set mycircle [list]
set myrect [list]
set myellipse [list]
set isellipse 1
for {set i 0} {$i < $length} {incr i} {
set value [::cv::contourArea [lindex $contours $i] 0]
if {$value > $max} {
set max $value
set maxindex $i
set contour [lindex $contours $i]
set bounding_rect [::cv::boundingRect $contour]
set mycircle [::cv::minEnclosingCircle $contour]
if {[catch {set myellipse [::cv::fitEllipse $contour]}]} {
set myrect [::cv::minAreaRect $contour]
set isellipse 0
}
}
}
::cv::drawContours $image1 $contours -1 [list 255 0 0 0] -1 $::cv::LINE_8 2 0 0
set x1 [lindex $bounding_rect 0]
set y1 [lindex $bounding_rect 1]
set x2 [expr $x1 + [lindex $bounding_rect 2]]
set y2 [expr $y1 + [lindex $bounding_rect 3]]
set color [list 255 255 0 0]
::cv::rectangle $image1 $x1 $y1 $x2 $y2 $color 1
set x1 [lindex $mycircle 0]
set y1 [lindex $mycircle 1]
set r [lindex $mycircle 2]
set color [list 0 255 255 0]
::cv::circle $image1 $x1 $y1 $r $color 1
set color [list 255 0 255 0]
if {$isellipse} {
set x1 [lindex $myellipse 0]
set y1 [lindex $myellipse 1]
set w [expr [lindex $myellipse 2]/2]
set h [expr [lindex $myellipse 3]/2]
set r [lindex $myellipse 4]
::cv::ellipse $image1 $x1 $y1 $w $h $r 0 360 $color 3
} else {
set box [::cv::boxPoints $myrect]
::cv::drawContours $image1 [list $box] -1 $color 3 $::cv::LINE_8 2 0 0
}
::cv::namedWindow "Display Image 3" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image 3" $image1
::cv::waitKey 0
::cv::destroyAllWindows
$image3 close
$image2 close
$image1 close
} on error {em} {
puts $em
}
Convex Hull test -
package require opencv
proc mysortproc {c1 c2} {
set value1 [::cv::contourArea $c1 0]
set value2 [::cv::contourArea $c2 0]
if {$value1 > $value2} {
return -1
} elseif {$value1 < $value2} {
return 1
} else {
return 0
}
}
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set image1 [::cv::imread $filename]
set skinColorUpper [list 15 204 153 0]
set skinColorLower [list 0 25 13 0]
# Try to filter
set hlsimage [cv::cvtColor $image1 $::cv::COLOR_BGR2HLS]
set rangeMask [cv::inRange $hlsimage $skinColorLower $skinColorUpper]
set blurred [cv::blur $rangeMask 10 10]
set image2 [::cv::threshold $blurred 200 255 $::cv::THRESH_BINARY]
$hlsimage close
$rangeMask close
$blurred close
set contours [::cv::findContours $image2 $::cv::RETR_EXTERNAL $::cv::CHAIN_APPROX_SIMPLE]
set scontours [lsort -command mysortproc $contours]
set contour [lindex $scontours 0]
set hull [cv::convexHull $contour 0 1]
for {set i 0} {$i < [llength $hull]} {incr i 2} {
set x [lindex $hull $i]
set y [lindex $hull [expr $i + 1]]
cv::circle $image1 $x $y 5 [list 255 0 0 0] 3
}
::cv::drawContours $image1 [list $hull] -1 [list 0 255 0 0] 1 $::cv::LINE_AA 2 0 0
set hull [cv::convexHull $contour 0 0]
set defects [::cv::convexityDefects $contour $hull]
foreach d $defects {
set startx [lindex $contour [expr [lindex $d 0] * 2]]
set starty [lindex $contour [expr [lindex $d 0] * 2 + 1]]
set endx [lindex $contour [expr [lindex $d 1] * 2]]
set endy [lindex $contour [expr [lindex $d 1] * 2 + 1]]
set farx [lindex $contour [expr [lindex $d 2] * 2]]
set fary [lindex $contour [expr [lindex $d 2] * 2 + 1]]
set depth [expr [lindex $d 3]/256]
cv::line $image1 $startx $starty $farx $fary [list 0 0 255 0] 2
cv::line $image1 $endx $endy $farx $fary [list 0 0 255 0] 2
cv::circle $image1 $farx $fary 5 [list 255 255 0 0] 3
}
::cv::namedWindow "Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Image" $image1
::cv::namedWindow "Check" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Check" $image2
::cv::waitKey 0
::cv::destroyAllWindows
$image2 close
$image1 close
} on error {em} {
puts $em
}
GeneralizedHoughBallard test -
package require opencv
try {
#
# Download files from:
# https://github.com/opencv/opencv/tree/master/samples/data
#
set img [::cv::imread "pic1.png" 0]
set templ [::cv::imread "templ.png" 0]
set hough [::cv::GeneralizedHoughBallard]
$hough setMinDist 100
$hough setLevels 360
$hough setDp 2
$hough setVotesThreshold 50
$hough setTemplate $templ
set result [$hough detect $img]
set pos [lindex $result 0]
set votes [lindex $result 1]
set img2 [::cv::cvtColor $img $::cv::COLOR_GRAY2BGR]
for {set i 0} {$i < [$pos rows]} {incr i} {
for {set j 0} {$j < [$pos cols]} {incr j} {
set x [expr int([$pos at [list $i $j] 0])]
set y [expr int([$pos at [list $i $j] 1])]
set scale [$pos at [list $i $j] 2]
set angle [$pos at [list $i $j] 3]
set h [expr int($scale * [$templ rows])]
set w [expr int($scale * [$templ cols])]
set rect [list $x $y $w $h $angle]
set box [::cv::boxPoints $rect]
set color [list 0 0 255 0]
::cv::drawContours $img2 [list $box] -1 $color 3
}
}
$pos close
$votes close
$hough close
$templ close
$img close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img2
::cv::waitKey 0
::cv::destroyAllWindows
$img2 close
} on error {em} {
puts $em
}
AGAST Algorithm for Corner Detection -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "blox.jpg"
try {
set img1 [::cv::imread $filename1 0]
set agast [::cv::AgastFeatureDetector]
set kp1 [$agast detect $img1]
set kpoint1 [::cv::drawKeypoints $img1 $kp1 None [list 255 0 0 0] \
$::cv::DRAW_MATCHES_FLAGS_DRAW_RICH_KEYPOINTS]
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $kpoint1
::cv::waitKey 0
::cv::destroyAllWindows
$kpoint1 close
$img1 close
} on error {em} {
puts $em
}
Brute-Force Matching with ORB Descriptors -
package require opencv
proc mysortproc {x y} {
set distance1 [lindex $x 3]
set distance2 [lindex $y 3]
if {$distance1 > $distance2} {
return 1
} elseif {$distance1 < $distance2} {
return -1
} else {
return 0
}
}
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set orb [::cv::ORB]
set result1 [$orb detectAndCompute $img1]
set result2 [$orb detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set bmatcher [::cv::BFMatcher $::cv::NORM_HAMMING 1]
set match [$bmatcher match $d1 $d2]
set match [lsort -command mysortproc $match]
set matches [lrange $match 0 10]
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $matches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$orb close
$bmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Flann-based descriptor matcher with BRISK Descriptors -
package require opencv
proc mysortproc {x y} {
set distance1 [lindex $x 3]
set distance2 [lindex $y 3]
if {$distance1 > $distance2} {
return 1
} elseif {$distance1 < $distance2} {
return -1
} else {
return 0
}
}
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set brisk [::cv::BRISK]
set result1 [$brisk detectAndCompute $img1]
set result2 [$brisk detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set fmatcher [::cv::FlannBasedMatcher FLANN_INDEX_LSH [list 6 12 1]]
set match [$fmatcher match $d1 $d2]
set match [lsort -command mysortproc $match]
set matches [lrange $match 0 10]
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $matches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$brisk close
$fmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Brute-Force Matching with ORB Descriptors and Ratio Test -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set orb [::cv::ORB]
$orb setWTA_K 3
set result1 [$orb detectAndCompute $img1]
set result2 [$orb detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set bmatcher [::cv::BFMatcher $::cv::NORM_HAMMING 0]
set matches [$bmatcher knnMatch $d1 $d2 2]
# Apply ratio test
set good [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.7 * $ndistance]} {
lappend good $m
}
}
}
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $good None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$orb close
$bmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Flann-based descriptor matcher with AKAZE Descriptors and Ratio Test -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set akaze [::cv::AKAZE]
set result1 [$akaze detectAndCompute $img1]
set result2 [$akaze detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set fmatcher [::cv::FlannBasedMatcher FLANN_INDEX_LSH [list 6 12 1]]
set matches [$fmatcher knnMatch $d1 $d2 2]
# Apply ratio test
set good [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.7 * $ndistance]} {
lappend good $m
}
}
}
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $good None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$akaze close
$fmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
ORB, Feature Matching+ Homography to find Objects -
package require opencv
proc mysortproc {x y} {
set distance1 [lindex $x 3]
set distance2 [lindex $y 3]
if {$distance1 > $distance2} {
return 1
} elseif {$distance1 < $distance2} {
return -1
} else {
return 0
}
}
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set orb [::cv::ORB]
set result1 [$orb detectAndCompute $img1]
set result2 [$orb detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set bmatcher [::cv::BFMatcher $::cv::NORM_HAMMING 1]
set match [$bmatcher match $d1 $d2]
set match [lsort -command mysortproc $match]
set dmatches [lrange $match 0 20]
set srcPts [list]
set dstPts [list]
for {set i 0} {$i < [llength $dmatches]} {incr i} {
set m [lindex $match $i]
set spoint [lindex $kp1 [lindex $m 0]]
set dpoint [lindex $kp2 [lindex $m 1]]
lappend srcPts [lindex $spoint 0] [lindex $spoint 1]
lappend dstPts [lindex $dpoint 0] [lindex $dpoint 1]
}
# Find homography matrix and do perspective transform
set src_pts [::cv::Mat::Mat 1 [expr [llength $srcPts]/2] $::cv::CV_32FC2]
$src_pts setData $srcPts
set dst_pts [::cv::Mat::Mat 1 [expr [llength $dstPts]/2] $::cv::CV_32FC2]
$dst_pts setData $dstPts
set MRes [::cv::findHomography $src_pts $dst_pts $::cv::RANSAC 5.0 2000 0.995]
set M [lindex $MRes 0]
set Mask [lindex $MRes 1]
if {![$M empty]} {
set h [$img1 rows]
set w [$img1 cols]
set pts [list 0 0 0 [expr $h-1] [expr $w-1] [expr $h-1] [expr $w-1] 0]
set dst [::cv::perspectiveTransform $pts $M]
::cv::polylines $img2 $dst 1 1 [list 255 255 255 0] 5
}
$M close
$Mask close
$src_pts close
$dst_pts close
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $dmatches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$orb close
$bmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
AKAZE, Feature Matching + Homography to find Objects -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set akaze [::cv::AKAZE]
set result1 [$akaze detectAndCompute $img1]
set result2 [$akaze detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set bmatcher [::cv::BFMatcher $::cv::NORM_HAMMING 0]
set matches [$bmatcher knnMatch $d1 $d2 2]
# Apply ratio test
set dmatches [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.7 * $ndistance]} {
lappend dmatches $m
}
}
}
set srcPts [list]
set dstPts [list]
foreach match $dmatches {
set spoint [lindex $kp1 [lindex $match 0]]
set dpoint [lindex $kp2 [lindex $match 1]]
lappend srcPts [lindex $spoint 0] [lindex $spoint 1]
lappend dstPts [lindex $dpoint 0] [lindex $dpoint 1]
}
# Find homography matrix and do perspective transform
set src_pts [::cv::Mat::Mat 1 [expr [llength $srcPts]/2] $::cv::CV_32FC2]
$src_pts setData $srcPts
set dst_pts [::cv::Mat::Mat 1 [expr [llength $dstPts]/2] $::cv::CV_32FC2]
$dst_pts setData $dstPts
set MRes [::cv::findHomography $src_pts $dst_pts $::cv::RANSAC 5.0 2000 0.995]
set M [lindex $MRes 0]
set Mask [lindex $MRes 1]
if {![$M empty]} {
set h [$img1 rows]
set w [$img1 cols]
set pts [list 0 0 0 [expr $h-1] [expr $w-1] [expr $h-1] [expr $w-1] 0]
set dst [::cv::perspectiveTransform $pts $M]
::cv::polylines $img2 $dst 1 1 [list 255 255 255 0] 5
}
$M close
$Mask close
$src_pts close
$dst_pts close
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $dmatches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$akaze close
$bmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
SIFT, Feature Matching + Homography to find Objects -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set sift [::cv::SIFT]
set result1 [$sift detectAndCompute $img1]
set result2 [$sift detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set fmatcher [::cv::FlannBasedMatcher FLANN_INDEX_KDTREE [list 5]]
set matches [$fmatcher knnMatch $d1 $d2 2]
# Apply ratio test
set dmatches [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.7 * $ndistance]} {
lappend dmatches $m
}
}
}
set srcPts [list]
set dstPts [list]
foreach match $dmatches {
set spoint [lindex $kp1 [lindex $match 0]]
set dpoint [lindex $kp2 [lindex $match 1]]
lappend srcPts [lindex $spoint 0] [lindex $spoint 1]
lappend dstPts [lindex $dpoint 0] [lindex $dpoint 1]
}
# Find homography matrix and do perspective transform
set src_pts [::cv::Mat::Mat 1 [expr [llength $srcPts]/2] $::cv::CV_32FC2]
$src_pts setData $srcPts
set dst_pts [::cv::Mat::Mat 1 [expr [llength $dstPts]/2] $::cv::CV_32FC2]
$dst_pts setData $dstPts
set MRes [::cv::findHomography $src_pts $dst_pts $::cv::RANSAC 5.0 2000 0.995]
set M [lindex $MRes 0]
set Mask [lindex $MRes 1]
if {![$M empty]} {
set h [$img1 rows]
set w [$img1 cols]
set pts [list 0 0 0 [expr $h-1] [expr $w-1] [expr $h-1] [expr $w-1] 0]
set dst [::cv::perspectiveTransform $pts $M]
::cv::polylines $img2 $dst 1 1 [list 255 255 255 0] 5
}
$M close
$Mask close
$src_pts close
$dst_pts close
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $dmatches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$sift close
$fmatcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
AffineFeature with KAZE, Feature Matching + Homography to find Objects -
package require opencv
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "box.png"
set filename2 "box_in_scene.png"
try {
set img1 [::cv::imread $filename1 0]
set img2 [::cv::imread $filename2 0]
set backend [::cv::KAZE]
set asift [::cv::AffineFeature $backend]
set result1 [$asift detectAndCompute $img1]
set result2 [$asift detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set matcher [::cv::FlannBasedMatcher FLANN_INDEX_KDTREE [list 5]]
set matches [$matcher knnMatch $d1 $d2 2]
# Apply ratio test
set dmatches [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.7 * $ndistance]} {
lappend dmatches $m
}
}
}
set srcPts [list]
set dstPts [list]
foreach match $dmatches {
set spoint [lindex $kp1 [lindex $match 0]]
set dpoint [lindex $kp2 [lindex $match 1]]
lappend srcPts [lindex $spoint 0] [lindex $spoint 1]
lappend dstPts [lindex $dpoint 0] [lindex $dpoint 1]
}
# Find homography matrix and do perspective transform
set src_pts [::cv::Mat::Mat 1 [expr [llength $srcPts]/2] $::cv::CV_32FC2]
$src_pts setData $srcPts
set dst_pts [::cv::Mat::Mat 1 [expr [llength $dstPts]/2] $::cv::CV_32FC2]
$dst_pts setData $dstPts
set MRes [::cv::findHomography $src_pts $dst_pts $::cv::RANSAC 5.0 2000 0.995]
set M [lindex $MRes 0]
set Mask [lindex $MRes 1]
if {![$M empty]} {
set h [$img1 rows]
set w [$img1 cols]
set pts [list 0 0 0 [expr $h-1] [expr $w-1] [expr $h-1] [expr $w-1] 0]
set dst [::cv::perspectiveTransform $pts $M]
::cv::polylines $img2 $dst 1 1 [list 255 255 255 0] 5
}
$M close
$Mask close
$src_pts close
$dst_pts close
set mcolor [list 255 0 0 0]
set scolor [list 0 0 255 0]
set match1 [::cv::drawMatches $img1 $kp1 $img2 $kp2 $dmatches None \
$mcolor $scolor $::cv::DRAW_MATCHES_FLAGS_NOT_DRAW_SINGLE_POINTS]
$d1 close
$d2 close
$asift close
$backend close
$matcher close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $match1
::cv::waitKey 0
::cv::destroyAllWindows
$match1 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Camera Calibration test -
package require opencv
set term [::cv::TermCriteria [expr $::cv::EPS | $::cv::COUNT] 30 0.001]
# prepare object points
set objp [::cv::Mat::zeros [expr 6*9] 3 $::cv::CV_32F]
for {set i 0} {$i < 9} {incr i} {
for {set j 0} {$j < 6} {incr j} {
$objp at [list [expr $i * 6 + $j] 0] 0 $j
$objp at [list [expr $i * 6 + $j] 1] 0 $i
$objp at [list [expr $i * 6 + $j] 2] 0 0
}
}
set objpoints [list]
set imgpoints [list]
set width 0
set height 0
#
# Download files from:
# https://github.com/opencv/opencv/tree/master/samples/data
# left01.jpg ~ left14.jpg (except left12.jpg)
#
set filenames [glob data/left*.jpg]
foreach file $filenames {
set img [::cv::imread $file]
set gray [cv::cvtColor $img $::cv::COLOR_BGR2GRAY]
set res [cv::findChessboardCorners $gray 6 9]
set ret [lindex $res 0]
set corners [lindex $res 1]
if {$ret == 1} {
cv::cornerSubPix $gray $corners 11 11 -1 -1 $term
set width [$img cols]
set height [$img rows]
cv::drawChessboardCorners $img 6 9 $corners $ret
cv::imshow "img" $img
cv::waitKey 500
lappend imgpoints $corners
lappend objpoints $objp
}
$img close
$gray close
}
$term close
set cameraMatrix [::cv::Mat::eye 3 3 $::cv::CV_64F]
set distCoeffs [::cv::Mat::zeros 5 1 $::cv::CV_64F]
set res [cv::calibrateCamera $objpoints $imgpoints $width $height $cameraMatrix $distCoeffs]
set ret [lindex $res 0]
set cameraMatrix [lindex $res 1]
set distCoeffs [lindex $res 2]
set rvecs [lindex $res 3]
set tvecs [lindex $res 4]
# Save to FileStorage
set fs [::cv::FileStorage]
$fs open "calibration.yaml" $::cv::FileStorage::WRITE
$fs writeMat "mtx" $cameraMatrix
$fs writeMat "dist" $distCoeffs
$fs close
# Undistortion
set img [::cv::imread "left12.jpg"]
set width [$img cols]
set height [$img rows]
set result [::cv::getOptimalNewCameraMatrix $cameraMatrix $distCoeffs \
$width $height 1 $width $height]
set newcameramtx [lindex $result 0]
set roix [lindex $result 1]
set roiy [lindex $result 2]
set roiw [lindex $result 3]
set roih [lindex $result 4]
set dst [::cv::undistort $img $cameraMatrix $distCoeffs $newcameramtx]
set dst2 [$dst rect $roix $roiy $roiw $roih]
::cv::imwrite "calibresult.png" $dst2
$dst close
$dst2 close
set R [cv::Mat::Mat 0 0 $::cv::CV_32F]
set rdst [::cv::initUndistortRectifyMap $cameraMatrix $distCoeffs $R $newcameramtx \
$width $height $::cv::CV_32F]
set mapx [lindex $rdst 0]
set mapy [lindex $rdst 1]
set dst [::cv::remap $img $mapx $mapy $::cv::INTER_LINEAR]
set dst2 [$dst rect $roix $roiy $roiw $roih]
::cv::imwrite "calibresult2.png" $dst2
$mapx close
$mapy close
$dst close
$dst2 close
$R close
$newcameramtx close
$img close
$cameraMatrix close
$distCoeffs close
foreach rvec $rvecs {
$rvec close
}
foreach tvec $tvecs {
$tvec close
}
$objp close
foreach imgp $imgpoints {
$imgp close
}
Pose Estimation, Render a Cube -
package require opencv
proc draw {img imgpts} {
set imgpts_c [list]
set imgpts_sub [list]
for {set i 0} {$i < 4} {incr i} {
set img_x [expr int([$imgpts at [list $i 0] 0])]
set img_y [expr int([$imgpts at [list $i 0] 1])]
lappend imgpts_sub $img_x $img_y
}
lappend imgpts_c $imgpts_sub
cv::drawContours $img $imgpts_c -1 [list 0 255 0 0] -3 $::cv::LINE_8 2 0 0
for {set i 0; set j 4} {$i < 4} {incr i; incr j} {
set img_x1 [expr int([$imgpts at [list $i 0] 0])]
set img_y1 [expr int([$imgpts at [list $i 0] 1])]
set img_x2 [expr int([$imgpts at [list $j 0] 0])]
set img_y2 [expr int([$imgpts at [list $j 0] 1])]
cv::line $img $img_x1 $img_y1 $img_x2 $img_y2 [list 255 0 0 0] 3
}
set imgpts_c [list]
set imgpts_sub [list]
for {set i 4} {$i < 8} {incr i} {
set img_x [expr int([$imgpts at [list $i 0] 0])]
set img_y [expr int([$imgpts at [list $i 0] 1])]
lappend imgpts_sub $img_x $img_y
}
lappend imgpts_c $imgpts_sub
cv::drawContours $img $imgpts_c -1 [list 0 0 255 0] 3 $::cv::LINE_8 2 0 0
}
# Load data
set fs [::cv::FileStorage ]
$fs open "calibration.yaml" $::cv::FileStorage::READ
set cameraMatrix [$fs readMat "mtx"]
set distCoeffs [$fs readMat "dist"]
$fs close
# prepare object points
set objp [::cv::Mat::zeros [expr 6*9] 3 $::cv::CV_32F]
for {set i 0} {$i < 9} {incr i} {
for {set j 0} {$j < 6} {incr j} {
$objp at [list [expr $i * 6 + $j] 0] 0 $j
$objp at [list [expr $i * 6 + $j] 1] 0 $i
$objp at [list [expr $i * 6 + $j] 2] 0 0
}
}
set axis [cv::Mat::Mat 8 3 $::cv::CV_32F]
$axis setData [list 0 0 0 0 3 0 3 3 0 3 0 0 \
0 0 -3 0 3 -3 3 3 -3 3 0 -3]
set term [::cv::TermCriteria [expr $::cv::EPS | $::cv::COUNT] 30 0.001]
set filenames [glob data/left*.jpg]
foreach file $filenames {
set img [::cv::imread $file]
set gray [cv::cvtColor $img $::cv::COLOR_BGR2GRAY]
set res [cv::findChessboardCorners $gray 6 9]
set ret [lindex $res 0]
set corners [lindex $res 1]
if {$ret == 1} {
cv::cornerSubPix $gray $corners 11 11 -1 -1 $term
set r [cv::solvePnP $objp $corners $cameraMatrix $distCoeffs]
set rvec [lindex $r 1]
set tvec [lindex $r 2]
set imgpts [cv::projectPoints $axis $rvec $tvec $cameraMatrix $distCoeffs]
draw $img $imgpts
cv::imshow "img" $img
cv::waitKey 500
$imgpts close
$rvec close
$tvec close
}
$corners close
$img close
$gray close
}
$term close
$objp close
$axis close
Epipolar Geometry -
package require opencv
proc drawlines {img1 lines pts1 pts2} {
set col [$img1 cols]
set row [$img1 rows]
for {set i 0} {$i < [$lines rows]} {incr i} {
set color [list [expr int(rand()*255)] [expr int(rand()*255)] [expr int(rand()*255)] 0]
set r0 [$lines at [list $i 0] 0]
set r1 [$lines at [list $i 0] 1]
set r2 [$lines at [list $i 0] 2]
set x0 0
set y0 [expr int(-$r2/$r1)]
set x1 $col
set y1 [expr int(-($r2+$r0*$col)/$r1)]
::cv::line $img1 $x0 $y0 $x1 $y1 $color 1
set p1 [$pts1 at [list $i 0] 0]
set p2 [$pts1 at [list $i 0] 1]
::cv::circle $img1 $p1 $p2 5 $color -1
}
}
#
# From https://github.com/opencv/opencv/tree/master/samples/data
#
set filename1 "left.jpg"
set filename2 "right.jpg"
try {
set img1 [::cv::imread $filename1 $::cv::IMREAD_GRAYSCALE]
set img2 [::cv::imread $filename2 $::cv::IMREAD_GRAYSCALE]
set akaze [::cv::AKAZE]
set result1 [$akaze detectAndCompute $img1]
set result2 [$akaze detectAndCompute $img2]
set kp1 [lindex $result1 0]
set d1 [lindex $result1 1]
set kp2 [lindex $result2 0]
set d2 [lindex $result2 1]
set bmatcher [::cv::BFMatcher $::cv::NORM_HAMMING 0]
set matches [$bmatcher knnMatch $d1 $d2 2]
$d1 close
$d2 close
$akaze close
$bmatcher close
set dmatches [list]
foreach match $matches {
foreach {m n} $match {
set mdistance [lindex $m 3]
set ndistance [lindex $n 3]
if {$mdistance < [expr 0.8 * $ndistance]} {
lappend dmatches $m
}
}
}
set srcPts [list]
set dstPts [list]
foreach match $dmatches {
set spoint [lindex $kp1 [lindex $match 0]]
set dpoint [lindex $kp2 [lindex $match 1]]
lappend srcPts [expr int([lindex $spoint 0])] [expr int([lindex $spoint 1])]
lappend dstPts [expr int([lindex $dpoint 0])] [expr int([lindex $dpoint 1])]
}
# Find homography matrix and do perspective transform
set src_pts [::cv::Mat::Mat 1 [expr [llength $srcPts]/2] $::cv::CV_32SC2]
$src_pts setData $srcPts
set dst_pts [::cv::Mat::Mat 1 [expr [llength $dstPts]/2] $::cv::CV_32SC2]
$dst_pts setData $dstPts
set result [::cv::findFundamentalMat $src_pts $dst_pts $::cv::FM_LMEDS 3.0 0.99]
set F [lindex $result 0]
set mask [lindex $result 1]
set nsrc_pts [cv::Mat::Mat 0 0 $::cv::CV_32SC2]
set ndst_pts [cv::Mat::Mat 0 0 $::cv::CV_32SC2]
for {set i 0} {$i < [$mask rows]} {incr i} {
set value [$mask at [list $i 0] 0]
# We select only inlier points
if {$value == 1} {
set srccol [$src_pts col $i]
$nsrc_pts push_back $srccol
set dstcol [$dst_pts col $i]
$ndst_pts push_back $dstcol
$srccol close
$dstcol close
}
}
$src_pts close
$dst_pts close
set colorimg1 [::cv::cvtColor $img1 $::cv::COLOR_GRAY2BGR]
set colorimg2 [::cv::cvtColor $img2 $::cv::COLOR_GRAY2BGR]
set lines1 [::cv::computeCorrespondEpilines $ndst_pts 2 $F]
drawlines $colorimg1 $lines1 $nsrc_pts $ndst_pts
::cv::imshow "imgag1" $colorimg1
set lines2 [::cv::computeCorrespondEpilines $nsrc_pts 1 $F]
drawlines $colorimg2 $lines2 $ndst_pts $nsrc_pts
::cv::imshow "imgag2" $colorimg2
::cv::waitKey 0
$nsrc_pts close
$ndst_pts close
$F close
$mask close
$lines1 close
$lines2 close
$colorimg1 close
$colorimg2 close
$img1 close
$img2 close
} on error {em} {
puts $em
}
Non-Photorealistic Rendering -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set img2 [::cv::detailEnhance $img]
set img3 [::cv::edgePreservingFilter $img2]
set img4 [::cv::stylization $img3]
::cv::imwrite "stylization.png" $img4
set penresult [::cv::pencilSketch $img3]
set img5 [lindex $penresult 0]
set img6 [lindex $penresult 1]
::cv::imwrite "pencilSketch_1.png" $img5
::cv::imwrite "pencilSketch_2.png" $img6
$img close
$img2 close
$img3 close
$img4 close
$img5 close
$img6 close
} on error {em} {
puts $em
}
High Dynamic Range Imaging test -
package require opencv
#
# From https://en.wikipedia.org/wiki/High-dynamic-range_imaging
# Download the four exposured images and test.
#
set times [list 0.0333 0.25 2.5 15.0]
set files [list 1.JPG 2.JPG 3.JPG 4.JPG]
set images [list]
try {
foreach f $files {
set img [::cv::imread $f]
lappend images $img
}
set a [::cv::AlignMTB]
set newimages [$a process $images]
for {set i 0} {$i < [llength $images]} {incr i} {
[lindex $images $i] close
}
set c [::cv::CalibrateDebevec]
set responseDebevec [$c process $newimages $times]
set mergeDebevec [::cv::MergeDebevec]
set hdrDebevec [$mergeDebevec process $newimages $times $responseDebevec]
::cv::imwrite "hdrDebevec.exr" $hdrDebevec
for {set i 0} {$i < [llength $newimages]} {incr i} {
[lindex $newimages $i] close
}
$responseDebevec close
set tonemapDrago [::cv::TonemapDrago 1.0 0.7 0.85]
set result [$tonemapDrago process $hdrDebevec]
set result2 [$result multiply 3]
set result3 [$result2 multiply 255]
::cv::imwrite "ldr-Drago.jpg" $result3
$tonemapDrago close
$result close
$result2 close
$result3 close
set tonemapMantiuk [::cv::TonemapMantiuk 2.2 0.85 1.2]
set result [$tonemapMantiuk process $hdrDebevec]
set result2 [$result multiply 3]
set result3 [$result2 multiply 255]
::cv::imwrite "ldr-Mantiuk.jpg" $result3
$tonemapMantiuk close
$result close
$result2 close
$result3 close
set tonemapReinhard [::cv::TonemapReinhard 1.5 0 0 0]
set result [$tonemapReinhard process $hdrDebevec]
set result2 [$result multiply 255]
::cv::imwrite "ldr-Reinhard.jpg" $result2
$tonemapReinhard close
$result close
$result2 close
$hdrDebevec close
} on error {em} {
puts $em
}
Stitcher test -
package require opencv
if {$argc < 1} {
puts "Please give an image list."
exit
}
try {
set imagelist [list]
for {set i 0} {$i < $argc} {incr i} {
set img [::cv::imread [lindex $argv $i] $::cv::IMREAD_COLOR]
lappend imagelist $img
}
set s [::cv::Stitcher $::cv::PANORAMA]
set result [$s stitch $imagelist]
$s close
::cv::imwrite output.png $result
$result close
} on error {em} {
puts $em
}
PCA example -
package require opencv
proc drawAxis {matrix x1 y1 x2 y2 color scale} {
set PI 3.1415926535897931
set angle [expr atan2($y1 - $y2, $x1 - $x2)]
set hypotenuse [expr sqrt(($y1 - $y2)*($y1 - $y2) + ($x1 - $x2)*($x1 - $x2))]
set qx [expr int($x1 - $scale * $hypotenuse * cos($angle))]
set qy [expr int($y1 - $scale * $hypotenuse * sin($angle))]
cv::line $matrix $x1 $y1 $qx $qy $color 1 $::cv::LINE_AA 0
set px [expr int($qx + 9 * cos($angle + $PI / 4))]
set py [expr int($qy + 9 * sin($angle + $PI / 4))]
cv::line $matrix $px $py $qx $qy $color 1 $::cv::LINE_AA 0
set px [expr int($qx + 9 * cos($angle - $PI / 4))]
set py [expr int($qy + 9 * sin($angle - $PI / 4))]
cv::line $matrix $px $py $qx $qy $color 1 $::cv::LINE_AA 0
}
proc getOrientation {matrix contour} {
set sz [expr [llength $contour] / 2]
set data_pts [::cv::Mat::Mat $sz 2 $::cv::CV_64F]
for {set i 0; set j 0} {$i < $sz} {incr i 1; incr j 2} {
$data_pts at [list $i 0] 0 [lindex $contour $j]
$data_pts at [list $i 1] 0 [lindex $contour [expr $j + 1]]
}
set pca [::cv::PCA $data_pts $::cv::DATA_AS_ROW]
set mean [$pca mean]
set cntr_x [expr int([$mean at [list 0 0] 0])]
set cntr_y [expr int([$mean at [list 0 1] 0])]
set eigenvalues [$pca eigenvalues]
set eigenvectors [$pca eigenvectors]
set eigen_vecs [list]
set eigen_val [list]
for {set i 0} {$i < 2} {incr i} {
set x [$eigenvectors at [list $i 0] 0]
set y [$eigenvectors at [list $i 1] 0]
lappend eigen_vecs $x $y
lappend eigen_val [$eigenvalues at [list $i 0] 0]
}
set p1_x [expr int($cntr_x + 0.02 * [lindex $eigen_vecs 0] * [lindex $eigen_val 0])]
set p1_y [expr int($cntr_y + 0.02 * [lindex $eigen_vecs 1] * [lindex $eigen_val 0])]
set p2_x [expr int($cntr_x - 0.02 * [lindex $eigen_vecs 2] * [lindex $eigen_val 1])]
set p2_y [expr int($cntr_y - 0.02 * [lindex $eigen_vecs 3] * [lindex $eigen_val 1])]
cv::circle $matrix $cntr_x $cntr_y 3 [list 255 0 255 0] 2
drawAxis $matrix $cntr_x $cntr_y $p1_x $p1_y [list 0 255 0 0] 1
drawAxis $matrix $cntr_x $cntr_y $p2_x $p2_y [list 255 255 0 0] 5
$mean close
$eigenvectors close
$eigenvalues close
$pca close
}
#
# Download from https://github.com/opencv/opencv/blob/master/samples/data/pca_test1.jpg
#
set filename "pca_test1.jpg"
try {
set src [::cv::imread $filename $::cv::IMREAD_COLOR]
::cv::namedWindow "Src" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Src" $src
set gray [::cv::cvtColor $src $::cv::COLOR_BGR2GRAY]
set bw [cv::threshold $gray 50 255 [expr $::cv::THRESH_BINARY | $::cv::THRESH_OTSU]]
set contours [cv::findContours $bw $::cv::RETR_LIST $::cv::CHAIN_APPROX_NONE]
for {set i 0} {$i < [llength $contours]} {incr i} {
set contour [lindex $contours $i]
set area [cv::contourArea $contour]
if {$area < 100 || $area > 100000} {
continue
}
cv::drawContours $src $contours $i [list 0 0 255 0] 2 $::cv::LINE_8 2 0 0
getOrientation $src $contour
}
$gray close
$bw close
::cv::namedWindow "Output" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Output" $src
::cv::waitKey 0
::cv::destroyAllWindows
$src close
} on error {em} {
puts $em
}
QRCodeDetector example -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img [::cv::imread $filename $::cv::IMREAD_COLOR]
set qrdetect [::cv::QRCodeDetector]
set result [$qrdetect detectAndDecode $img]
puts "Decode result: [lindex $result 0]"
set box [lindex $result 1]
set code [lindex $result 2]
set m [$box rows]
set n [$box cols]
if {[$box empty] != 1} {
for {set i 0} {$i < $m} {incr i} {
for {set j 0} {$j < $n} {incr j} {
set x1 [expr int([$box at [list $i $j] 0])]
set y1 [expr int([$box at [list $i $j] 1])]
set x2 [expr int([$box at [list $i [expr ($j+1)%$n]] 0])]
set y2 [expr int([$box at [list $i [expr ($j+1)%$n]] 1])]
cv::line $img $x1 $y1 $x2 $y2 [list 255 128 0 0] 5
}
}
}
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img
if {[$code empty] != 1} {
::cv::namedWindow "Display QR code image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display QR code image" $code
}
::cv::waitKey 0
::cv::destroyAllWindows
$box close
$code close
$qrdetect close
$img close
} on error {em} {
puts $em
}
People detection by using HOGDescriptor -
package require opencv
# The example file can be downloaded from:
# https://github.com/opencv/opencv/blob/master/samples/data/vtest.avi
set filename "vtest.avi"
set v [::cv::VideoCapture file $filename]
if {[$v isOpened]==0} {
puts "Open Video file $filename failed."
exit
}
set hog [cv::HOGDescriptor 64 128 16 16 8 8 8 8 9]
$hog setSVMDetector [$hog getDefaultPeopleDetector]
set hog_d [cv::HOGDescriptor 48 96 16 16 8 8 8 8 9 1 -1 0.2 0 64 0]
$hog_d setSVMDetector [$hog getDaimlerPeopleDetector]
set mode 0
while {[$v isOpened]==1} {
try {
set frame [$v read]
set t [::cv::getTickCount]
if {$mode == 0} {
set rects [$hog detectMultiScale $frame 0 8 8 0 0 1.05 2 0]
} else {
set rects [$hog_d detectMultiScale $frame 0 8 8 0 0 1.05 2 1]
}
set t [expr [cv::getTickCount] - $t]
if {$mode == 0} {
set buf "Mode: Default ||| FPS: [format "%0.1f" [expr [::cv::getTickFrequency] / $t]]"
} else {
set buf "Mode: Daimler ||| FPS: [format "%0.1f" [expr [::cv::getTickFrequency] / $t]]"
}
::cv::putText $frame $buf 10 30 $::cv::FONT_HERSHEY_PLAIN 2.0 [list 0 0 255 0] 2
set length [llength $rects]
for {set i 0} {$i < $length} {incr i} {
set rlist [lindex $rects $i]
# Slightly shrink the rectangles to get a nicer output.
set width [lindex $rlist 2]
set height [lindex $rlist 3]
set x1 [expr int([lindex $rlist 0] + $width * 0.1)]
set y1 [expr int([lindex $rlist 1] + $height * 0.07)]
set x2 [expr int($x1 + $width * 0.8)]
set y2 [expr int($y1 + $height * 0.8)]
set color [list 0 255 0 0]
::cv::rectangle $frame $x1 $y1 $x2 $y2 $color 2
}
::cv::imshow "People detector" $frame
$frame close
set key [::cv::waitKey 1]
if {$key==[scan "q" %c] || $key == 27} {
break
} elseif {$key == [scan " " %c]} {
if {$mode == 0} {
set mode 1
} else {
set mode 0
}
}
} on error {em} {
break
}
}
$hog close
$hog_d close
$v close
::cv::destroyAllWindows
The following program shows how to detect faces in an image (using CascadeClassifier).
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img1 [::cv::imread $filename $::cv::IMREAD_COLOR]
set img2 [::cv::cvtColor $img1 $::cv::COLOR_BGR2GRAY]
set img3 [::cv::equalizeHist $img2]
$img2 close
#
# From https://github.com/opencv/opencv/tree/master/data/lbpcascades
# For test ::cv::CascadeClassifier command
#
set xmlfile "lbpcascades/lbpcascade_frontalface.xml"
set classifier [::cv::CascadeClassifier $xmlfile]
set rects [$classifier detectMultiScale $img3]
set length [llength $rects]
for {set i 0} {$i < $length} {incr i} {
set rlist [lindex $rects $i]
set x1 [lindex $rlist 0]
set y1 [lindex $rlist 1]
set x2 [expr $x1 + [lindex $rlist 2]]
set y2 [expr $y1 + [lindex $rlist 3]]
set color [list 255 0 0 0]
::cv::rectangle $img1 $x1 $y1 $x2 $y2 $color 2 $::cv::LINE_AA 0
}
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img1
::cv::waitKey 0
::cv::destroyAllWindows
$classifier close
$img1 close
$img3 close
} on error {em} {
puts $em
}
Below is an DNN example (face detector) -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img1 [::cv::imread $filename $::cv::IMREAD_COLOR]
set blob [::cv::dnn::blobFromImage $img1 1.0 300 300 \
[list 104 117 123 0] 1 1]
# From https://github.com/opencv/opencv/tree/master/samples/dnn
# model is from models.yml and config is from face_detector folder's file
# Just download to test
set model "res10_300x300_ssd_iter_140000.caffemodel"
set config "deploy.prototxt"
set net [::cv::dnn::readNet $model $config Caffe]
$net setPreferableBackend $::cv::dnn::DNN_BACKEND_DEFAULT
$net setPreferableTarget $::cv::dnn::DNN_TARGET_CPU
$net setInput $blob
set data [$net forward]
set msize [$data size]
for {set i 0} {$i < [lindex $msize 2]} {incr i} {
set confidence [$data at [list 0 0 $i 2] 0]
if {$confidence > 0.9} {
set startX [expr int([$data at [list 0 0 $i 3] 0] * [$img1 cols])]
set startY [expr int([$data at [list 0 0 $i 4] 0] * [$img1 rows])]
set endX [expr int([$data at [list 0 0 $i 5] 0] * [$img1 cols])]
set endY [expr int([$data at [list 0 0 $i 6] 0] * [$img1 rows])]
cv::rectangle $img1 $startX $startY $endX $endY [list 255 0 0 0] 1
}
}
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img1
::cv::waitKey 0
::cv::destroyAllWindows
$net close
$blob close
$img1 close
} on error {em} {
puts $em
}
Object Detection using YOLOv3 -
package require opencv
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
try {
set img1 [::cv::imread $filename $::cv::IMREAD_COLOR]
set blob [::cv::dnn::blobFromImage $img1 [expr 1.0/255.0] 416 416 \
[list 0 0 0 0] 1 0]
#
# Load names of classes
#
set labelsFile [open "coco.names"]
set classes [list]
while {1} {
set line [gets $labelsFile]
if {[eof $labelsFile]} {
close $labelsFile
break
}
lappend classes $line
}
#
# Download the model from: https://pjreddie.com/darknet/yolo/
#
set model "yolov3.weights"
set config "yolov3.cfg"
set net [::cv::dnn::readNet $model $config Darknet]
$net setPreferableBackend $::cv::dnn::DNN_BACKEND_OPENCV
$net setPreferableTarget $::cv::dnn::DNN_TARGET_CPU
$net setInput $blob
# Runs the forward (pass to the output layers)
set names [$net getUnconnectedOutLayersNames]
set outputdata [$net forwardWithNames $names]
# Scan through all the bounding boxes output from the network
set boxes [list]
set confidences [list]
set classIDs [list]
foreach mat $outputdata {
for {set i 0} {$i < [$mat rows]} {incr i} {
set rowdata [$mat row $i]
set scores [$rowdata colRange 5 [$mat cols]]
set valueloc [::cv::minMaxLoc $scores]
set confidence [lindex [lindex $valueloc 1] 0]
set classId [lindex [lindex $valueloc 1] 1]
if {$confidence > 0.5} {
set centerX [expr int([$rowdata at [list 0 0] 0] * [$img1 cols])]
set centerY [expr int([$rowdata at [list 0 1] 0] * [$img1 rows])]
set width [expr int([$rowdata at [list 0 2] 0] * [$img1 cols])]
set height [expr int([$rowdata at [list 0 3] 0] * [$img1 rows])]
set left [expr $centerX - $width/2]
set top [expr $centerY - $height/2]
lappend classIDs $classId
lappend confidences $confidence
lappend boxes [list $left $top $width $height]
}
$scores close
$rowdata close
}
$mat close
}
# Perform non maximum suppression to eliminate redundant overlapping boxes
set indicates [::cv::dnn::NMSBoxes $boxes $confidences 0.5 0.3]
foreach ind $indicates {
set box [lindex $boxes $ind]
set x1 [lindex $box 0]
set y1 [lindex $box 1]
set x2 [expr $x1 + [lindex $box 2]]
set y2 [expr $y1 + [lindex $box 3]]
set color [list 255 178 0 0]
::cv::rectangle $img1 $x1 $y1 $x2 $y2 $color 3
set class [lindex $classes [lindex $classIDs $ind]]
append class ": " [format %.6f [lindex $confidences $ind]]
set thickness 2
set tsize [::cv::getTextSize $class $::cv::FONT_HERSHEY_SIMPLEX 0.5 $thickness]
set twidth [lindex $tsize 0]
set theight [lindex $tsize 1]
set baseline [expr [lindex $tsize 2] + $thickness]
::cv::rectangle $img1 $x1 [expr $y1-$theight-$baseline] \
[expr $x1+$twidth] [expr $y1] \
[list 255 255 255 0] -1
::cv::putText $img1 $class $x1 [expr $y1-$baseline] \
$::cv::FONT_HERSHEY_SIMPLEX 0.5 [list 0 0 255 0] $thickness
}
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img1
::cv::waitKey 0
::cv::destroyAllWindows
$net close
$blob close
$img1 close
} on error {em} {
puts $em
}
Text Detection (EAST text detector) -
package require opencv
proc decode {scores geometry scoreThresh} {
set size [$scores size]
set height [lindex $size 2]
set width [lindex $size 3]
set confidences [list]
set boxes [list]
for {set y 0} {$y < $height} {incr y} {
for {set x 0} {$x < $width} {incr x} {
# Extract the scores (probabilities)
set score [$scores at [list 0 0 $y $x] 0]
if {$score < $scoreThresh} {
continue
}
set x0 [$geometry at [list 0 0 $y $x] 0]
set x1 [$geometry at [list 0 1 $y $x] 0]
set x2 [$geometry at [list 0 2 $y $x] 0]
set x3 [$geometry at [list 0 3 $y $x] 0]
set angle [$geometry at [list 0 4 $y $x] 0]
set offsetX [expr $x * 4.0]
set offsetY [expr $y * 4.0]
set cosA [expr cos($angle)]
set sinA [expr sin($angle)]
set h [expr $x0+$x2]
set w [expr $x1+$x3]
set ox [expr $offsetX + $cosA * $x1 + $sinA * $x2]
set oy [expr $offsetY - $sinA * $x1 + $cosA * $x2]
set p1x [expr $ox - $sinA * $h]
set p1y [expr $oy - $cosA * $h]
set p3x [expr $ox - $cosA * $w]
set p3y [expr $oy + $sinA * $w]
set box [list [expr ($p1x+$p3x)*0.5] [expr ($p1y+$p3y)*0.5] $w $h \
[expr -$angle*180.0/3.1415926535897932384626433832795]]
lappend confidences $score
lappend boxes $box
}
}
return [list $boxes $confidences]
}
if {$argc != 1} {
exit
}
set filename [lindex $argv 0]
set SIZE 320
try {
set img1 [::cv::imread $filename $::cv::IMREAD_COLOR]
set blob [::cv::dnn::blobFromImage $img1 1.0 $SIZE $SIZE \
[list 123.68 116.78 103.94 0] 1 0]
#
# Text detection model: https://github.com/argman/EAST
# Download the model from:
# https://www.dropbox.com/s/r2ingd0l3zt8hxs/frozen_east_text_detection.tar.gz?dl=1
#
set model "frozen_east_text_detection.pb"
set net [::cv::dnn::readNet $model]
$net setPreferableBackend $::cv::dnn::DNN_BACKEND_OPENCV
$net setPreferableTarget $::cv::dnn::DNN_TARGET_CPU
$net setInput $blob
# Runs the forward (pass to the output layers)
set names [list "feature_fusion/Conv_7/Sigmoid" "feature_fusion/concat_3"]
set out [$net forwardWithNames $names]
set scores [lindex $out 0]
set geometry [lindex $out 1]
# Process the output
set result [decode $scores $geometry 0.5]
set boxes [lindex $result 0]
set confidences [lindex $result 1]
set hRatio [expr [$img1 rows]/double($SIZE)]
set wRatio [expr [$img1 cols]/double($SIZE)]
# Get the final predictions
set indicates [::cv::dnn::NMSBoxes $boxes $confidences 0.5 0.4]
foreach ind $indicates {
set box [lindex $boxes $ind]
# Use ::cv::boxPoints to get RotatedRect 4 points
set box2 [::cv::boxPoints $box]
set mybox [list]
# Scale the bounding box coordinates based on the respective ratios
for {set i 0} {$i < 8} {incr i 2} {
set x [expr int([lindex $box2 $i] * $wRatio)]
set y [expr int([lindex $box2 [expr $i + 1]] * $hRatio)]
lappend mybox $x $y
}
# Draw the box
#::cv::drawContours $img1 [list $mybox] -1 [list 0 255 0 0] 2
for {set i 0} {$i < 8} {incr i 2} {
set x1 [lindex $mybox [expr $i%8]]
set y1 [lindex $mybox [expr ($i+1)%8]]
set x2 [lindex $mybox [expr ($i+2)%8]]
set y2 [lindex $mybox [expr ($i+3)%8]]
::cv::line $img1 $x1 $y1 $x2 $y2 [list 0 255 0 0] 2
}
}
$scores close
$geometry close
::cv::namedWindow "Display Image" $::cv::WINDOW_AUTOSIZE
::cv::imshow "Display Image" $img1
::cv::waitKey 0
::cv::destroyAllWindows
$net close
$blob close
$img1 close
} on error {em} {
puts $em
}