The complexity of the target tracking environment results in a large number of false alarms and missed detections in the input of the tracking system. Furthermore, phenomena such as track crossing exist in multi-target tracking problems, posing significant challenges to the initiation, maintenance, and termination of tracks in the tracking system. Tracking algorithms can typically eliminate outliers, use historical observation information to fill in missing track information, smooth trajectories, predict the number of targets, add batch information of targets, etc. The performance of tracking mainly depends on the effectiveness of data association.