The analysis of tennis recognition model for human health based on computer vision and particle swarm optimization algorithm

The study aims to solve the problem of tennis picking for players in the training process and realize intelligent tennis picking. An intelligent tennis picking robot is studied to recognize and position tennis balls. First, the tennis recognition algorithm based on HSV (hue, saturation, value) color space is used to identify the tennis ball, and the coordinates of tennis and obstacles are obtained by background difference and OF (optical flow). Second, particle swarm optimization (PSO) that has excellent global planning ability and support vector machine (SVM) that has good obstacle avoidance performance are applicable because there may be some obstacles in tennis courts. Therefore, the traditional PSO and SVM are combined to obtain the optimized PSO. And the simulation comparison experiment is carried out on the Matlab simulation software. Finally, the model is tested and 50 random screenshots of tennis videos collected on the spot, and tennis photos downloaded on the network are tested in the dataset. The results show that the number of tennis balls correctly identified by the proposed algorithm is 248 and that of tennis balls wrongly identified is 8. Its recognition accuracy is 96.88% and the time spent is 9.33 s. The algorithm proposed provides some ideas to solve the problem of tennis picking for tennis players.