| People’s health has always been a livelihood issue of great concern to the Party and the Country.The Country has issued the outline of "Healthy China 2030" as the guiding ideology for building a healthy china in all regions.At the same time,people are increasingly concerned about their physical health.Therefore,national physical health monitoring has become a necessary means to understand the level of national physical health.In view of the time consuming and low detection efficiency of conventional physical health monitoring equipment,there is a wide range of intelligent upgrading of physical health monitoring detection methods.The main research of this paper is to upgrade the pull-up detection method based on action recognition technology,identify the movements of the tester through Kinect detection system,judge whether the movements are standard and count.At the same time,considering the difficulty of pull-up and the inability of many test subjects to complete a standard movement,push-up is introduced to evaluate the upper body strength of test subjects.Similarly,based on motion recognition technology,the detection method of push-up has been intelligently upgraded.The main contents of this paper are as follows:(1)The action recognition model is established on base of the node triangle method.Motion recognition models were established for pull-ups and push-ups respectively,and the elbow angle was defined by the node-triangle method.Different angle eigenvalues correspond to different recognition actions,and the hidden Markov model of the actions was established according to the angle characteristics of the actions,which provided the basis for the following Kinect motion recognition.(2)The pull-up counting method is presented on base of the action recognition.A detection system was built based on Kinect sensor to collect the coordinate data of the head node in the process of movement.According to the periodic variation characteristics of the data itself,it was determined to count the changes of the head coordinates by recording them.In addition,by setting the horizontal bar coordinate and angle threshold to judge whether the pull-up is standard.(3)The push-up counting method is presented on base of the action recognition,and the evaluation of upper body strength is presented.Analyze the elbow angle data in the process of push-up movement,and determine the count by recording the elbow angle changes.In addition,determine the body posture through the transformation of quaternion,and detect whether the push-up is standard through the body posture.Finally,the independent weight method was used to weight the two test items of 30 students,and the initial upper body strength score was obtained.After that,the BMI index was introduced to compare the upper body strength scores among students with different BMI levels,and appropriate correction parameters were formulated for the thin and fat students respectively,so as to exclude the influence of weight on the upper body strength test.Through the study of the above issues,this paper successfully used Kinect sensor to complete the visual detection of pull-ups and push-ups,providing an efficient and accurate detection method for physical fitness monitoring,and improved the evaluation method of upper body strength through BMI index. |
