Skip to content

PawarMukesh/German-Traffic-Sign-Detction

Repository files navigation

GERMAN TRAFFIC SIGN DETECTION

Detect The German Traffic Sign Using YoloV5

DATA COLLECTION:

Data Collected form Rubixe For Educational Internship

TASK: OBJECT DETECTION

DATA INFORMATION:

  • Total 4290 traffic sign images are present in 39 classes Each class contain 110 images
  • Split data with the help of splitfolder library.
  • 88 Images for training and 22 images for validation
  • Create bounding boxes with the help of label-img tool and makesense.ai website

German Traffic Signs

image

DATA PREPRATION:

  • Prepare folder structure that can be accept by YoloV5.
  • Total 3438 images for training and 857 images for validation present in 39 classes.
  • Create a bounding boxes with the help of label-img And makesense.ai website according to YoloV5.

STEPS TO USE YOLOV5

  • Cloning the YoloV5 file from official repository.
  • Changing the directory of yolov5
  • Installing the dependencies
  • Download all versions pre-trained weights

STEPS BEFORE TRAINING CUSTOM DATASET:

  1. Go to yolov5/data/

  2. Open coco128.yaml

  3. Edit the following inside it:

    A. Training and Validation file path

    B. Number of classes and Class names.

TRAINING YOLOV5 MODEL

  • Set images size 128 with batch of 8
  • Train model on 50 epochs
  • Gives the data file path as well as give pre-trained weights path.

VISUALISE THE TRAINING METRICS WITH THE HELP OF TENSORBOARD

AFTER TRAINING THE MODEL

image

VALIDATION IMAGES PREDICTION:

image

PREDICTED IMAGES:

image

image

CHALLENGES FACED:

  • Facing problem to understand the business case.
  • challenge faced in bounding boxes creation
  • Assign same no for all classes
  • Made mistake in yolov5 folder structure
  • Take lots of time to create bounding boxes

WHAT WE LEARN:

  • Convert classification task to object detection to improve skill in object detection
  • Understand the YoloV5 folder structure as well as learn label-img tool.
  • Learn pytorch library.