Three-dimensional human body action reconstruction system and method and action training system

Abstract

The invention discloses a three-dimensional human body action reconstruction system and method and an action training system. The method comprises: capturing an action image of the target human body through the corrected binocular vision-based binocular camera; carrying out two-dimensional human body articulation point identification on the action image on a two-dimensional image; reconstructing three-dimensional articulation points of the target human body based on binocular stereoscopic vision and the two-dimensional human body articulation points. Therefore, the three-dimensional action ofthe target human body is reconstructed. A depth map is generated without complex graphic transformation; errors caused by a depth change process can be avoided. According to the human body three-dimensional action recognition method, the situation that objects around the target human body are mistakenly recognized as one part of the target human body is avoided, the requirement for an applicationscene is lowered, and the technical problems that an existing depth camera human body three-dimensional action recognition method is high in requirement for the application scene and low in recognition accuracy in a complex application scene are solved.

Description

technical field [0001] The present application relates to the technical field of motion recognition, in particular to a three-dimensional human motion reconstruction system, method and motion training system. Background technique [0002] With the development of artificial intelligence, human body posture recognition technology has made important breakthroughs on the basis of larger data sets and powerful computing capabilities available in the era of big data. [0003] The existing 3D human motion recognition method is to use a depth camera to capture the 3D motion of the human body, and then evaluate whether the motion meets the standards. The method of using the depth camera to capture the 3D motion of the human body uses near-infrared light to cover and irradiate the view range to be tested, and Calculate the time difference or phase difference between near-infrared light emission and emission, and convert the distance of each point of the target to be measured, so as to...

AI Technical Summary

Problems solved by technology

[0004] The purpose of this application is to provide a 3D human body motion reconstruction system, method and motion training system, which are used to solve the existing depth camera human body 3D motion recognition method, which has high requirements for application scenarios and accurate recognition in complex application scenarios Low rate of technical issues

Images