Design And Implementation Of Golf Teaching Assistant Analysis System Based On Target Tracking Technology

Along with the rapid development of Internet technology,the golf intelligent system that can accurately evaluate and feedback the user's action posture presents great potential.However,current golf smart products lack systematic guidance and the ability to regulate user actions.Moreover,the analysis of human motion by the existing intelligent systems on the market mostly relies on three-dimensional depth information,and the products have such the disadvantages as high equipment cost,large acquisition of data by environmental interference,and poor real-time performance of algorithms for reconstructing three-dimensional information.Therefore,it is of great significance to design an efficient and professional golf teaching aid analysis system.In this paper,the target monocular camera is used to track the collected golf moving images,and the extracted human motion features are modeled and compared,and the user posture is analyzed and evaluated in real time.The specific technology includes the following three points:(1)A target tracking method based on pixel features is proposed.Based on the fast moving speed of the golf club and the light and small target characteristics of the shaft,the foreground and background regions are automatically updated by using the target motion trajectory information,and the image foreground is extracted and processed by the adaptive double threshold three-frame difference method.On this basis,using the pixel features of the foreground target,three auxiliary algorithms of rod-body region determination,motion region detection and pixel feature density are introduced,and the goal of the club is tracked in real time through cyclic progression.Finally,an area tracking algorithm based on the direction of the target vector is designed to correct the club positioning features to ensure the integrity and standardization of subsequent feature output.(2)A real-time evaluation method for golf swing posture comparison is proposed.In order to solve the problem of the direction of human joint movement,this paper proposes the feature of the joint rotation angle as the attitude evaluation.Firstly,for the two key angles of the left arm and the arm,which are the most important in the golf action,the distance similarity measure is used to make a rough evaluation of the user's swing posture.Secondly based on the analysis of the golf side-swing video,the swing plane in the user's motion is monitored to complement the attitude comparison method,which makes the golf attitude comparison algorithm more complete.Finally,in order to evaluate the global attitude action,this paper proposes a weighted direction rotation angle similarity measurement method,which makes the golf attitude comparison algorithm more applicable;(3)A golf teaching aid analysis system is designed.Based on the above technical research and development results,the golf teaching aid analysis system was completed.Firstly,using the sequence image captured by the monocular camera,the human joint information is extracted by the attitude estimation algorithm,and the two-dimensional motion trajectory of the club is obtained by the target tracking algorithm.Secondly by using the key motion and posture characteristics of golf,the overall motion of the user is evaluated.Finally,the golf teaching aid analysis prototype system designed and implemented in this paper is carried out in real scenes.The results show that the system has achieved good results in real-time and accuracy,and has high practical value.