Moving Target Detection Algorithm Based on Improved YOLOv10

This paper proposes a moving target detection algorithm based on improved YOLOv 10. Alarge number of target motion videos are shot with high-definition industrial cameras, and then the YOLOv10 model is used to detect moving targets. The coordinate attention mechanism is proposed to improve the original PSA attention mechanism, and a dynamic upsampling strategy is introduced in the neck network. Experiments have shown that this method can accurately identify moving targets in videos and greatly improve the accuracy.