Research And Implementation Of Sports Teaching System Based On Kinect

The Olympic motto is Faster, Higher and Stronger. The athletes always challengethe record and themselves. With the successfully hostingof the Beijing OlympicGames in2008, more and more people realize that the physical exercise is beneficialto physical and mental health. The concept of popular sports is accepted by thecountrymen gradually. We also reach a consensus on the concept that sports power isbuilt on the basis of the improvement of the whole people diathesis. The improvementof the whole people diathesis relies on the professional sports guidance. However, inour nation, due to lack of the teaching for popular sports and professional guidance,the competitive level of many sports enthusiasts is difficult to improve, or evenserious sports injuries and other issues have happened. Therefore, it’s urgent todevelop a low price and excellent popular sports teaching system.The popular sports teaching system should have the function of motion capture,reconstruction and analysis. At present, there are many studies on the motion capture,while the equipment is expensive and more motion constraints. In addition, very fewabout human motion reconstruction and comparison analysis, and yet there is no agood solution. To address the above problems, in this paper, the research work iscarried out as follows:Firstly, in the three-dimensional human motion capture aspect, this topic use thelatest Microsoft Kinect depth information detection device to measure the key nodedata of human skeleton and convert it to motion data that match with the hierarchicalstructure of human skeleton. Kinect uses the advanced technology of optical encodingtechnology in the process of information acquisition. The technique does not requirethe special light-sensitive chip, only an ordinary CMOS sensor chip can put it intoaction, which makes the cost of depth information acquisition is decreased greatly. Inthe motion data generation, Kinect uses the motion data which match with thehierarchical structure of human skeleton. One hand, the amount of data to be logged isdecreased greatly, on the other hand, it provides convenience to the subsequentmotion construction and analysis.Secondly, in the reconstruction of three-dimensional human motion, this topicreconstructs the three-dimensional motion of the human body vividly with the technology of skinned mesh animation and individuality manikin models. Accordingto the different application, two human motion reconstruction methods are designed inthis paper, namely the motion reconstruction method which based on the standardhuman models and the individuality manikin models. Two reconstruction methods canobserve action sequence or a fixed posture from an arbitrary viewing point in theOpenGL environment.Finally, in the aspect of three-dimensional motion analysis, in this paper, aimingat the practical problems existing in the motion data, we designed a variety ofcomparison analysis methods and conducted the confirmatory test. This part is mainlyto reduce the error of the manually contrast, thus gives more intelligent quantitativeevaluation result.