针对当前运动员训练的需求,结合当前的计算机视觉技术,提出一种运动员错误动作识别的模型.对体育运动错误动作三维建模的原理进行分析,然后结合错误动作识别的流程,以南拳中的高空腾飞作为研究对象,用三维坐标构建其动作形式化描述规则,并采用贝叶斯算法对错误动作进行识别.最后对上述的方案进行验证,得到本文构建的贝叶斯算法在检测的准确率等方面都比传统的要高,以此验证了贝叶斯在对南拳错误动作中的有效性.%In view of the needs of current athletes training and combined with the current computer vision technology, a model based on the error recognition of athletes is proposed. The principle of sports error action 3D modeling is analyzed. Combined with the wrong action recognition process, taking high altitude Nanquan as the research object, the 3D coordinates of the athletes are applied to build the formal description of action rules, and the Bayesian algorithm is used to identify the wrong action. Finally the above scheme is verified. The Bayesian algorithm has higher detection accuracy than the traditional ones.
展开▼