RecognitionTrafficLigth

Traffic light recognition

yolo must be installed, if not, follow the instructions indicated in: https://learnopencv.com/ultralytics-yolov8/#How-to-Use-YOLOv8?

pip install ultralytics

also supervision of Roboflow must be installed

pip install supervision

also must be installed the usual modules in computer vision: numpy, cv2

All needed modules can be installed with a plain pip

Once downloaded to disk, execute:

RecognizeTrafficLights.py

References:

https://medium.com/@mayank.siddharth/image-segmentation-lightning-fast-yolov8-unleashed-no-gpu-required-with-a-sleek-6-7mb-model-7854cb29c54d

https://roboflow.com/how-to-filter-detections/yolov8

#https://blog.roboflow.com/convert-bboxes-masks-polygons/

Leo Ueno. (Aug 15, 2023). Ultimate Guide to Converting Bounding Boxes, Masks and Polygons. Roboflow Blog: https://blog.roboflow.com/convert-bboxes-masks-polygons/

https://medium.com/thedeephub/object-tracking-with-mean-shift-and-cam-shift-algorithms-using-opencv-fc2f30327199

https://github.com/ikramnoun/self-driving-car (video downloaded from)