Sports Evaluation System Based on BlazePose and Depth Camera
In the current automated sports testing and training,subjects often need to wear a variety of sensors,which affects the performance.On the other hand,it is difficult to describe the human motion posture comprehensively and accurately only relying on two-dimensional images or depth cameras.Aiming at the above problems,this paper proposes a motion evaluation system suit-able for conventional physical measurement projects.Firstly,the 2d human pose estimation model BlazePose is combined with the depth camera to achieve the accurate positioning of key points of human pose in 3d space,and then the track of key points is smoothed and estimated by Kalman filter.On this basis,according to the thought of finite state machine,pull-ups,for example,the evaluation method for multi-index cervix projects design is completed to fully detect all kinds of fouls,experiments show that the overall detection rate reaches 89%,correct processing speed up to 20 frames per second,which can be used in the actual test and play a guiding role for personal training.
key points of human postureaction evaluationsports testingdepth camera
万方数据
维普
