Research on Traffic Police Area Target Detection and Dynamic Tracking Based on YOLOv8 and Optical Flow Method

With the rapid development of deep learning technologies in the field of autonomous driving, significant progress has been made in traffic police gesture detection and recognition. However, traffic police area detection still faces certain challenges and requires further optimization and innovation. To address issues such as low accuracy and poor robustness in dynamic traffic police area recognition, this paper designs a traffic police area tracking system tailored to the demands of real-world traffic scenarios. In urban road conditions, YOLOv8 is used for detecting the traffic police area, and corner detection algorithms from optical flow methods are employed for position tracking, improving the system’s effectiveness in traffic environments.