Analysis Of Tennis Serve Behavior Based On Video Image Processing

Human motion analysis based on video is an important research direction in the computer vision field.It is to detect the moving objects from the video sequence,extract the key parts of the human body,obtain useful information of the human body movement,and realize the further analysis and recognition of human motion and posture.The traditional teaching mode of tennis serve has been deeply rooted in the present teaching.For athletes,in order to master the serve technology skillfully,they must practice repeatedly based on the coach's guidance for a long time.Long-term use of experience-oriented teaching methods or the coach's subjective awareness of the athletes technical guidance and supervision have seriously hampered the improvement of teaching level.The author take the tennis flat serve as the research background based on full understanding of the relevant research work.The paper's main research contents are as follows:(1)In order to improve the accuracy and training efficiency of flat serve in tennis,the joint points of the serve arm are marked with color,and the serve video of the tennis ball is collected by the high-speed camera in this paper.The coordinates of the joints are replaced by the coordinates of the markers in each frame.In the process of video processing,the dictionary is reconstructed from a series of noise images.The non-interfering serve images are reconstructed by the sparse representation idea,and the mixed Gaussian background modeling is used to extract the motion foreground.After obtaining the foreground of motion,the marker points are extracted by color feature,and binarization is carried out.Then,the contours of the marker points are searched,and the coordinates of the circle are taken as the coordinates of the joint.Experimental data show that the success rate of serve can be 97.24% when the ratio of hitting point and height in the range of [1.37,1.44].(2)In order to further study the trajectory of mark points,this paper will take the serve trajectory of shoulder mark point as the research object and establish a tennis service model based on improved support vector machine.Aiming at the problem that the traditional support vector machine does not solve the penalty parameter c and the parameter value ? of the kernel function,the classification accuracy is taken as the fitness function and the particle swarm algorithm is used to improve it.After training the model with training data,the optimal penalty parameter c and the kernel parameter ? are returned.The experimental results show that the model can accurately classify the serve trajectory with the accuracy of 97.5%.Then the model is compared with the traditional SVM.The classification error of PSO_SVM is much smaller than SVM.