Research on Recognition of Football Trajectory Based on Robot Vision System

In the video tracking system, to analyze the movement of the football, it is necessary to know the movement position of the football at each moment. In this way, it is necessary to identify and separate the target object-football in each frame of the video sequence. Determine its location. In order to effectively detect and track moving targets in football videos, it is necessary to study the algorithm of target detection and tracking in football video. The currently used algorithm has the problem of poor detection and tracking of moving targets in dynamic scenes. The football video image sequence can be segmented based on the cylindrical distance according to the main color, and the segmentation result can be morphologically processed to remove the holes and burrs generated in the segmentation process. Mark other connected regions in the field area as candidate targets. The method adopted in this paper is a robot vision system, which can improve the accuracy and flexibility of the robot system. Therefore, it is of great value to explore the robot vision system and control strategy.