The research describes a system it use adaptive background reduction and object tracking to count vehicles on the highways and roads. Background subtraction that adapts, picture vehicle counting, segmentation, and vehicle tracking are its four primary phases of operation. Video streams from stationary cameras placed on roads and highways are used. This system's precise vehicle tracking and counting is meant to improve traffic surveillance. Its capacity to efficiently count automobiles and produce extremely accurate findings are among its key qualities. The study analyses traffic recordings for vehicle detection, classification, and counting using sort algorithms based on artificial intelligence. Vehicles in three classes-cars, trucks, buses-were identified and categorised with an average accuracy of using YOLOv5 and YOLOv8, a deep learning technique. After detection, the number of passing vehicles was counted using the SORT algorithm, which yielded an average accuracy rate of 95%. Based on the experimental results, it is possible to successfully recognise, classify, and count vehicles in flowing traffic films by utilising the YOLOv5 and YOLOv8 to SORT algorithms.
Comparative Analysis Yolov5 and Yolov8 for Vehicle Classification
06.12.2024
1856219 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch