Research On Human Health Monitoring Method Based On Depth Camera

With the continuous improvement of the living standards of the people and the aging of the population,people are paying more and more attention to the health status of their motor functions,and the decline of motor function is the main factor that jeopardizes people's daily life,especially the motor function of the knee joint.Bad changes will directly affect people's normal activities.Aiming at this phenomenon,this paper proposes a human body health monitoring method based on depth camera,and develops a human health monitoring system,which effectively improves the practicability and convenience of human knee health monitoring method through non-contact monitoring.The main work and innovations of this paper are as follows:(1)A deep learning method for obtaining human joint points in depth images by DenseNet is proposed.A human joint point data set composed of depth images is marked,and training and testing are performed on the data set,reaching PCK@0.4 of 97.99%performance indicator.(2)The noise reduction of the human joint point data obtained by the Kinect camera is obtained,and the motion waveforms of the knee joint of the human body are obtained firom the stereoscopic spatial data of the human motion process,and the low frequency motion signal is retained while suppressing the high frequency by the wavelet decomposition processing.The noise signal is reconstructed by wavelet reconstruction to obtain the motion waveform of the knee joint after noise reduction.(3)The active and passive motion models of the human body are established.The active motion model is to monitor the health of the knee joint by the subject's autonomously performing the prescribed action.The passive motion model is exercised by the subject following the healthy simulation model.The health monitoring result of the knee joint is obtained by analyzing the degree of fit between the tester's motion process and the healthy simulation model.The two monitoring results were combined to obtain an evaluation of the knee joint health status of the subject.(4)A human health monitoring system is built to realize the collection,processing and health prediction of human knee joint motion data.The design of the visual interface makes the operation of the system more convenient.At the same time,the system can display the human knee joint in motion.Changes in the process and real-time monitoring of k:nee health.Through the method proposed in this paper,the non-contact monitoring of human knee joint health can be effectively realized,and people can provide more convenient health monitoring programs.The monitoring results can reflect the health status of human knee joint motor function.