Research on Traffic Vehicle Target Tracking Method Based on YOLO Algorithm and ResNet

Aiming at the unsatisfactory accuracy of traffic vehicle tracking in the field of intelligent traffic, the study proposes a traffic vehicle target tracking method based on You Look only Once algorithm and residual network. The method firstly uses the improved You Look only once algorithm to detect and locate the vehicle target, and in order to further improve the target detection accuracy, the study introduces the attention mechanism, the decoupling head strategy and the feature fusion layer to improve the detection algorithm. Afterwards, the study applies Kalman filter and residual network to achieve the trajectory tracking of the target. The experimental findings indicate that the suggested method's average detection accuracy for identifying and locating various target kinds is 87.48%. During the tracking process, the average tracking speed of the proposed method is 12.07 FPS, the average tracking success rate is 98.52%, and the average tracking accuracy is 93.78%. The proposed traffic vehicle target tracking method is able to achieve better vehicle target tracking accuracy in different complex scenarios while ensuring higher tracking speed. In the area of intelligent transportation, the suggested technique can lessen traffic congestion and enhance the intelligence of autonomous driving even further.