短跑运动图像人体关节点离自动识别系统的研制与应用-运动人体科学专业毕业论文.docxVIP

结合短跑运动规律用于确定人体各关节点的区域跟踪；针对各关节点的特异性和各时相 结合短跑运动规律用于确定人体各关节点的区域跟踪；针对各关节点的特异性和各时相 的特异性，对不同时相的各关节点采取不同的跟踪区域中心和半径，并根据动作连贯性 进行自适应调整。根据圆心到圆周的各点距离相等这一原理，找到与关节点周围轮廓线 上特征点距离相等的点(一般为中点)作为所求关节点。 本研究根据运动的连续性，判断识别人体关节点出现特殊情况的运动图像。并利用 环节长法，由邻近关节点的坐标及环节斜率计算进入人体轮廓的可见侧关节点；利用像 素差异，提取可见侧手臂轮廓解决可见侧手臂关节点进入躯干轮廓的问题；将可见侧关 节点坐标移动半个周期，计算获得遮挡侧对应关节点坐标数据；利用像素值判断手位于 髋关节前后来解决手对髋关节确定的影响。综合应用轮廓线识别与灰度识别两种方法， 以轮廓线上特征点确定体内关节点为主，以灰度差异识别为辅，二者相互补充，提高了 自动识别精度和效率。 本研究研发出了可在Windows系统平台运行的、用Visual C++6．0编制的“短跑 人体关节点的自动识别软件系统计算机应用程序，实现了计算机自动识别短跑运动图 像中的人体关节点，使运动录像解析系统准确及时地反馈运动信息并具有良好的可操作 性。 关键字：短跑人体关节点自动识别计算机 Iv AbstractDevelopment Abstract Development of modern science and technology has promoted sports science and technology progress tremendously，high—tech has applied to motion training broadly，which improved training horizontal，moving engineering level and competition achievement．In train with the successive steps tllat development of sports science and motion，the improvement of the science training method and monitoting a means moves improved the effective training，importance improving motion achievement，already becoming one of sports developing direction．That science moves training is to build one kind of motion training pattern based on training intensity and control taking that enginery monitors the index and the technology monitoring an index as feedback inforillation．At present，an training effectuation to monitor main application’s from the technology monitors is the move analytical system of picture recording． But，moving analytical system of picture recording is returning the grave defect back to existence，being that the unable timely feedback information and operation are complicated mainly．Be one or two monthly after the image being sh()t about three moons ability gets the testing data even，can not realize real time supervisory control，the gains data not having reality guiding value．But affect a kinematics studying the bottleneck that the fleetness couples back depends on the data acquisition handles