Application of Computerized 3D Human Model Training Technology in Sports Movement Correction

In order to improve the ability of judging and correcting wrong movements in sports training, a 3d modeling and detection method of bad movements based on computer-aided 3d human body model feature decomposition is proposed. The motion behavior is analyzed by using machine vision image processing method, and the image acquisition model of motion error actions is constructed. The multi-dimensional wavelet scale decomposition method is used to decompose the collected motion error action images at pixel level, and the gray outline model of motion error actions is constructed, and the three-dimensional feature quantity of motion error actions is extracted. The three-dimensional modeling detection of motion errors is realized according to the three-dimensional human model feature decomposition method. The simulation results show that the accuracy of the 3d modeling detection of sports error actions is high, the detection error rate is low, and the real-time correction capability of the error actions is strong.