Human motion posture correction method and system based on computer vision

Abstract

The invention relates to a human motion posture correction method and system based on computer vision, and the method comprises the steps: obtaining the posture information of a user through videos recorded by a plurality of cameras from a plurality of angles, carrying out the comparison through combining with standard motions, discovering a non-standard position of a user motion, and giving guidance. A user can obtain intelligent sports skill guidance without wearing extra equipment. According to the method, the convolutional neural network model in deep learning is used for detecting the positions of the human body key points captured by the camera in the image, the precision can be improved by training the model, and the precision required by action guidance can be achieved. And meanwhile, the positions of the plurality of cameras can be set according to the actual environment, so that compared with the method of using kinect equipment, the method has a better capturing effect on various complex actions of the user.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a method and system for correcting human movement posture based on computer vision. Background technique [0002] Fitness or learning motor skills is a highly professional sport. If the body movements are not done correctly or the amount of exercise is too large or too small, it will have a greater impact on its effect, and may even bring about the opposite effect. For example, negative consequences such as muscle strains or joint injuries may occur. At present, there are few electronic devices that can provide relatively complete guidance for users to exercise or learn sports skills. Most of them use electronic devices with built-in acceleration sensors, gravity sensors and other sensors to record the user's heart rate, exercise duration, distance and other data. To guide the user's exercise duration, exercise intensity, etc. There are also sensors includin...

AI Technical Summary

Problems solved by technology

This existing technology has the following defects: the Kinect device uses structured light technology to measure the depth of field, so as to determine the spatial position of joint points. Due to the limitation of the structured light method itself, it can only achieve better results indoors and when the distance between people is relatively close. , for people at a slightly longer distance, its accuracy is greatly reduced

And due to the limitations of the kinect itself, it only has a good detection effect when the person is facing the device, and cannot recognize the complex movements of the person.

And this method does not have a good description of the time difference between the action to be tested and the standard action, such as the action is done faster or slower, or the person's action is slightly delayed compared to the standard action, etc.

Images