This repository contains the code for deploying and setting up plant leaf Image Detection, Image Segmentation, and Disease Detection models deployed on the Debian AI Transforma boards.
Model Image Detection, Image Segmentation, disease detection demo on live video.
newest_output_video.mp4
Green bounding boxes means healthy, red means disease. Corresponding colored masks are also applied onto leaves, making it easier to detect which ones have disease.
The YOLO model detects the leaves, returning the detected area into the image segmentation model. The segmentation model processes the image, leaving only areas with the leave intact and making the rest of the image black, feeding it into the disease detection model which returns the plant's disease.
For the visualization, we overlay the detected bounding boxes, image masks (red if diseased, green if healthy), and sometimes display the name of the disease. Below our results on different types of images.