Human body action automatic scoring method based on RGBD camera

Abstract

The invention discloses a human body action automatic scoring method based on an RGBD camera. The human body action automatic scoring method comprises the following steps: (1) acquiring joint point coordinate data; (2) performing data processing of joint vector coordinates; (3) calculating the similarity between the standard frame and the test frame; (4) performing time sequence similarity measurement; (5) performing automatic scoring. In the medical field, the movement posture of a patient with dyskinesia is subjected to rehabilitation training according to the standard of a normal person, and rehabilitation of the body of the patient is facilitated; in sports events, dance training and other occasions, the problem that manual action scoring lacks objectivity or is not uniform in scale can be effectively solved, the action of the sportsman is compared with the action synthesized in the standard database, scoring operation is conducted, and an automatic scoring method is formed.

Description

technical field [0001] The invention relates to an automatic scoring method for human body movements, in particular to an automatic scoring method for human body movements based on an RGBD camera. Background technique [0002] With the development of information technology, the research on human motion has attracted the attention and investment of many institutions at home and abroad. The U.S. Department of Defense project led by the Massachusetts Institute of Technology is dedicated to motion recognition and monitoring in video surveillance. Domestic research on human movement started relatively late, but has also achieved good results. Sign language research conducted by the Institute of Computing Technology, Chinese Academy of Sciences has achieved a high recognition rate. [0003] The RGBD cameras that have emerged in recent years can use depth information to obtain real-time human joint coordinates in 3D space, providing more forms and directions for human motion resea...

AI Technical Summary

Problems solved by technology

However, these joint points that make up the reference plane are not exactly on the same plane, and there is a high possibility of large distortion, resulting in the failure of posture correction

In his master's thesis "Research and Application of Human Action Recognition Based on Kinect Skeleton Information", Sui Longfei proposed to calibrate the coordinate system of human body posture and camera coordinate system through rotation and translation to realize the normalization of motion. This normalization method Only the difference between the two coordinate systems is eliminated, and the influence of the difference between different moving individuals on the similarity calculation is not considered

Images