Road signs detection using YOLO v5 object detection algorithm with PyTorch

This work presents to detect road signs in a few seconds for avoid accidents. For this work, there is utilized YOLO v5 object detection algorithm with PyTorch. In this method, image preprocessing is used the acquired image and further various augmentation is used to increase the accuracy of the system in the worst case such as blur, noise in the image, placed angle and so on. YOLO v5 algorithm makes system less complex to design and provides better accuracy, with this The training time in algorithm reduces from about ~30hrs to just ~2 hrs for 2,800 images. The designed algorithm gives an accuracy of 94% for four class sign defined.