Determining Method for Potential Injury Attitude Based on Three-dimensional Image Analysis

摘要

Potential injured area of athletes by image analysis may produce a frontier technology. In the method for determining the attitude of sports injuries, the potential injury-prone parts such as the wrist, waist and knees and other more hidden location during image capture almost can not guarantee the continuity of image acquisition, resulting in discontinuous movement in image features, which is hard to form reasonable judgment. Three-dimensional positioning method of motion magnitude is proposed on basis of ML, jointing three-dimensional dynamic model of the target, readily obtaining information of radial distance, azimuth and Doppler of athlete according to visual data, establishing the maximum likelihood estimation on moving amplitude in three-dimensional target positioning, and achieve judgment of the target. Simulation results show that the potential sports injuries pose based on three-dimensional image analysis can effectively improve judging accuracy of injury attitude, internal division of media space distribution of injury, reducing injury probability in the long course and providing better protection for potential movement.