PREDICTION OF THE BEST HIT POINT OF TENNIS SERVE BASED ON MEAN SHIFT ALGORITHM

The prediction accuracy of traditional tennis serve best hitting point prediction method is greatly affected by external factors such as the complexity of the environment, resulting in low prediction accuracy and efficiency. To address the above problems, the best hitting point prediction method of tennis serve based on the Mean Shift algorithm is studied. After acquiring and processing the video image of tennis serve, we detect the moving object in the image and track the tennis ball movement in the image by using the Mean Shift algorithm. Finally, the Kalman filter is used to predict the best hitting point of the serve. The experimental results demonstrate that the studied prediction method is efficient, improves the prediction accuracy by at least 50% on average, and can be applied in practice to improve the training effect of athletes.