A human posture estimation method based on OpenPose and Kinect and a rehabilitation training system

Abstract

The invention discloses a human posture estimation method based on OpenPose and Kinect and a rehabilitation training system. The system comprises a human posture estimation module based on OpenPose and Kinect, a scene interactive rehabilitation training virtual scene module and a three-dimensional joint motion trajectory database. The human posture estimation module based on OpenPose and Kinect fuses the depth information of Kinect and 2D human joints obtained by OpenPose algorithm to obtain 3D human joints. The scene interactive rehabilitation virtual scene part builds the progressive rehabilitation virtual scene based on Unity3D platform, and realizes the virtual agent's motion control and other functions. The three-dimensional joint motion trajectory database is used to store and read the motion trajectory data of each joint in the process of rehabilitation training. The system can capture the spatial coordinates of three-dimensional human joints in real time, scene interactive rehabilitation training is lively and interesting, and can provide targeted rehabilitation training according to the patient's situation.

Description

technical field [0001] The invention belongs to the intersection technology of the field of computer vision and rehabilitation robot, and relates to a human body posture estimation method and a rehabilitation training system based on OpenPose and Kinect. Background technique [0002] Since the 1990s, robot-assisted rehabilitation training technology has developed rapidly and attracted the attention of developed countries. Existing research and clinical applications have shown that rehabilitation training robots can provide safe, reliable, targeted and adaptive rehabilitation training for patients with limb motor dysfunction caused by stroke and spinal cord injury, and can improve limb motor dysfunction. It is of great significance to improve the quality of rehabilitation training for patients, promote early recovery of patients, and reduce the burden on families and society. [0003] In recent years, in order to develop more extensive rehabilitation training methods and fur...

AI Technical Summary

Problems solved by technology

However, the bone binding algorithm that comes with Kinect is very susceptible to the influence of lighting, foreground occlusion and human body self-occlusion, and it cannot be recognized or misidentified.

Due to partial limb disability, stroke patients often need to wear some support or fixed equipment to maintain body balance. As a result, the scene interaction system based on Kinect's built-in bone binding algorithm cannot be applied to rehabilitation robots that partially block the patient's body.

Moreover, most of the existing rehabilitation training scenario interaction systems do not provide goal-oriented virtual game scenarios for rehabilitation training according to the different disability states of patients' limbs and different rehabilitation stages, so that they cannot provide targeted rehabilitation training according to the patient's rehabilitation situation. Program

Images