Deep learning based image recognition for autonomous driving traffic (dataset: https://www.bdd100k.com)
A fined YOLOV5 model with custom dataset. Find more details in https://medium.com/@wumeng.maggie/hands-on-real-time-traffic-light-detection-train-your-own-dataset-with-yolov5-e5ca56188bef
To batch reading the local image file through CV2 library. Read in the cordinantes and preprocess to YOLOV5 format.