Research On Non-contact Monitoring Method Of Human Knee Joint Movement State

With the aggravation of the aging of the society in China,the problem of human health has become more and more prominent and gradually attracted attention.As the saying goes,"old people feet first decline",the health status of the knee joint is closely related to people’s motor function and thus affects their daily life.The traditional wearable assessment of knee joint health has some problems,such as dependence on medical equipment,complexity of wearability and difficulty of operation by non-medical personnel.Aiming at the defects of traditional monitoring methods,it is of application value and research significance to use optical principle to extract key information and establish knee joint motion model to carry out non-contact technology research on knee joint health monitoring.The main work and innovation of the thesis are as follows:(1)The knee joint motion image data set was built based on the depth camera,the main human joints and auxiliary points of the depth image were marked in the data set through the self-designed annotation software,and the main joint model of the human body with the addition of the auxiliary points of the quadrieps was constructed by using the neural network fusion of Densenet and A2 J.The accuracy of94.03% at PCK@0.2 was obtained.(2)On the basis of the existing human hip-knee-ankle joint motion model,an improved model with the additional of quadriceps auxiliary points was proposed;the Kalman filter algorithm was used to perform real-time noise reduction on the extracted knee joint related motion information Interference treatment;based on the three medical scoring scales of HSS,AKS,and Lysholm,a non-contact specific knee joint scoring standard and health indicators with additional auxiliary points are designed.(3)Completed the design of the non-contact human knee health monitoring system,including the software and hardware research and development of the system equipment,and provided two monitoring modes for preoperative prevention and postoperative rehabilitation.By testing the monitoring system on volunteers,the traditional inertial measurement unit(IMU)is compared with the first,The results showed that the non-contact system can monitor the motion state highly consistent with the real situation;secondly,the evaluation effect of the traditional medical knee joint scoring scale was compared,and the evaluation result of the non-contact system reached a performance with an error of 5.95%,which finally verified the practical value of the non-contact human knee health monitoring system.The research method of this thesis completes the interactive,all-round,and full-cycle effective monitoring of human knee joint health.The monitoring system has good robustness and effectiveness,and realizes the non-contact monitoring method of knee joint health and the goal of simple and convenient knee joint health management.