Sports Video Evaluation and Three-dimensional Human Motion Simulation of Teenagers in Physical Exercise

In recent years, teenagers’ physical exercise has become more and more important. In pattern recognition, intelligent human-machine interface, virtual reality and other aspects, video technology is the current research focus. Video analysis methods have good application prospects in intelligent monitoring, sports analysis, animation generation, etc. In the experimental part, the new approach of motion target selection based on dynamic background construction put forward in this paper was compared and analyzed with background subtraction method, time difference method and optical flow method in terms of recognition time consumption and false separation rate. 0.95%-1.03% was the fluctuation range of the misclassification rate of the method in this paper, and 1.64%-1.76% was the fluctuation range of the misclassification rate of the background subtraction method; 2.85%-3.06% was the fluctuation range of mis-separation rate of time difference method, and 4.02%-4.16% was the fluctuation range of mis-separation rate of optical flow method. It was not difficult to see that this method can better separate the motion images in sports video.