Exercise training evaluation method based on virtual environment

Abstract

The invention belongs to the field of somatosensory interaction and particularly provides an exercise training evaluation method based on a virtual environment. The method comprises the following steps that 1, a virtual teaching platform is established; 2, exercise trajectories are calculated; 3, exercise training is evaluated. The virtual teaching platform designed by adopting Unity 3D uses skeleton points acquired by Kinect V2 as a basis, keeps 30 frames per second to acquire standard vectors of standard actions and coordinates of spine points and calculates heights of standard action persons. By comparing the heights of persons participating in training with the heights of the standard action persons, offset coordinates of the spine points are determined, spatial exercise curves of standard actions are calculated according to the skeleton lengths of the persons participating in training, and accordingly continuous exercises are identified and graded.

Description

Technical field [0001] The invention belongs to the field of somatosensory interaction, and specifically is a sports training evaluation method based on a virtual environment. Background technique [0002] Most of the current sports training requires on-site instruction and teaching. The training cost is too high, or you can follow the learning after watching the video. After the end, you cannot self-correct irregular movements. CN 106485055 B proposes a method of combining angle and distance to recognize actions, but this method takes the spine point as the contrast center, and only sets some actions to compare at the end node, which cannot reflect the spatial motion curve of continuous motion. CN 107213619 A proposes a sports training evaluation method, which completes the automatic determination and guidance of sports through the combination of software and hardware, which reduces the workload of the coach, but does not propose a specific mathematical recognition method. Summ...

AI Technical Summary

Problems solved by technology

[0002] Most of the current sports training needs on-site guidance and teaching, the training cost is too high, or follow the learning after watching the video, and cannot self-correct irregular movements after the end

CN 106485055 B proposes a method combining angle and distance to recognize actions, but this method takes the spine point as the center of comparison, and only sets some actions at the end nodes for comparison, which cannot reflect the spatial motion curve of continuous motion

CN 107213619 A proposes a sports training evaluation method, through the combination of software and hardware to complete the automatic judgment and guidance of sports, which reduces the workload of the coach, but does not propose a specific mathematical identification method

Images