Optimized YOLOv4 Algorithm for Car Detection in Traffic Flow

The vehicle detection accuracy and actual in images videos appear to be very tough critical duties a key technology traffic system. Specifically, under convoluted conditions. As result, the presented study proposes single-stage deep neural networks YOLOv4-3L, YOLOv4-2L, YOLOv4-GB, YOLOv3-GB. After optimizing network structure by adding more layers right positions with amount of filters, dataset will repaired noise reduced before being sent mentoring. This research applied YOLOv3 YOLOv4. In this OA-Dataset is collect used, data set manually labeled care different weathers scenarios, as well for end-to-end training network. Around same time, optimized YOLOv4 demonstrate significant degree 99.68 % precision 91 %. speed algorithm are found higher than that previous algorithms.