Research on target detection and tracking algorithm of football game video based on deep learning

Sports programs are popular with the public because of their unique attraction. Therefore, it is of great significance and economic value to analyze and study the videos in sports events. Taking a football match as an example, this paper studies how to effectively detect and track moving objects on the court. Therefore, it is very necessary to recognize and track the moving objects in football. The existing methods are not ideal for detecting and tracking moving objects in dynamic scenes. On this basis, combined with Kalman filtering technology, a target tracking algorithm based on TMHI is designed. On this basis, using Kalman filtering algorithm, the position of the target mass center and the rectangular area around the target are obtained when the goal is scored, and the target tracking is realized. Experiments show that this algorithm can effectively detect and track moving targets in football videos.