Intelligent Recognition Method of Human Motion Global Features Based on Kinect Skeleton Information

摘要

In order to solve the problems of low recognition accuracy and poor real-time performance in the existing training and recognition of human movement posture under complex background, an intelligent recognition method of human motion global features based on Kinect skeleton information is designed. The data of skeletal pose feature points is obtained from Kinect, and the pose feature vector is calculated. The current real-time pose feature vector is matched with the preset standard pose feature vector. The matching degree of the two postures is fed back in real time to achieve the purpose of pose recognition. The experimental results show that the method can effectively match and recognize human action posture, with high recognition accuracy and strong real-time performance Scalability.