Realization Of Human Gait Analysis And Center Of Gravity Estimation Method Based On Kinect

The gait analysis and center of gravity estimation process are to obtain multiple gait parameter information of people while walking through scientific methods and estimate the center of gravity of the person in dynamic sports.It is of great significance in the fields of rehabilitation medicine,sports.Kinect can extract the position information of the human body joints and locate the joint points by extracting the person's outline in the field of view,thereby obtaining the human body posture and position.Based on the analysis of gait characteristics and the basic principles of center of gravity estimation,this thesis studies the fundamental technologies and implementation methods of gait analysis and center of gravity estimation based on Kinect.The main research work is as follows:(1)Given the lack of data and mutations in the Kinect human skeleton model,this article uses the Vicon human body model's joint point compression to establish the ViconKinect skeleton mapping model as the standard of the original skeleton model.It uses cubic spline interpolation.To achieve data synchronization between the original model and the mapping model,the Savitzky Golay filter algorithm is used to reduce the mutation of the original model data and compensate for the missing joint point data to improve the stability and accuracy of the data the Kinect,human skeleton model.(2)Aiming at the problem that Kinect's original skeleton information is challenging to use in gait analysis directly,and the data reference value is limited,this paper studies the calculation methods of joint angle,gait phase and space-time parameters.First,by analyzing the characteristics of human joint motion and the sensor's principle,the joint angle can be calculated directly using adjacent joint points in space.Secondly,through the comparative analysis of human gait cycle and foot and knee joint information,the method of gait phase division based on gait feature threshold and machine learning algorithm is realized,and the advantages and disadvantages of the two methods are compared to verify The result and error of the two.Finally,combining walking kinematics and human gait characteristics,the calculation of gait time and space parameters is realized.(3)Aiming at the problem of the inability to directly estimate the human center of gravity based on Kinect,this paper studies the calculation method of the human center of gravity and the enveloping area of the center of gravity.First,by analyzing the characteristics of human joint motion and sensor principles,based on kinematics and segmented human body structure model,a preliminary estimation of the human body's centre of gravity is achieved.Secondly,starting from the actual use of Kinect,taking into account the deviation of the proximal and distal ends of the body segment,a Kinect-based human center of gravity estimation correction model is established.Finally,through the analysis of the moving boundary points,the human center of gravity's envelope area is calculated.This thesis combines human motion characteristics and Kinect sensors to study and realize human gait analysis and center of gravity estimation methods,including skeleton processing,gait algorithm,and gravity model centre.Finally,the experimental results verify that the parameters are effective and available.