The Research Of Sports Video Object Tracking Based On Hybrid Algorithm

Sports training require continuous improvement with the increased level of sports competition. In the past, the coaches just intuitively assess the training, this approach has failed to meet the training requirements of competitive sports now. Therefore, computer vision technology is adopted. Machine vision has better accuracy and memory than human vision. It can quickly capture the moving target and record the movement data and provide more scientific data description of the action.Researching on the characteristics of sports video, the paper proposed an hybrid non-rigid target tracking method based on the mean shift algorithm and color histogram algorithm. Mean shift algorithmThe mean shift algorithm is based on the non-parametric kernel density estimation theory, using gradient method to calculate the extreme points of the probability density function. The algorithm has features of no parameters, fast pattern matching.Color histogram algorithm can simultaneously estimate the position of target object and the approximate shape. So it has excellent performance in tracking non-rigid object in sports video.In order to test the effectiveness of the hybrid algorithm, the paper simulate sports video detection and tracking with the MATLAB platform. In the test, the novel method does not track lost and test results demonstrate the effectiveness, adaptability of the new method.