基于图像分割的网球发球动作误差识别方法

Error Recognition Method of Tennis Serve Action Based on Image Segmentation

针对传统动作误差识别方法存在识别误差大的问题,提出基于图像分割的网球发球动作误差识别方法。建立网球发球动作图像的特征检测模型,分割网球发球图像的多尺度特征。通过自适应学习和尺度变换方法,进行网球发球动作视频采集图像的联合特征点定位和模糊动作特征检测,采用关联谱挖掘和细节特征分解方法,构建网球发球动作视频采集图像的多尺度检测模型,实现对网球发球动作视频采集图像的误差识别。仿真结果表明,采用该方法进行网球发球动作误差识别的精度较高,对错误动作的可靠性识别性能较好。

Aiming at the problem of large recognition error in traditional action error recognition methods,a tennis serve action error recognition method based on image segmentation is proposed.The feature detection model of tennis serve action image is established.The multi-scale features of tennis serve image are segmented.After changing the adaptive learning and scale transformation method,the joint feature point location and fuzzy action feature detection of tennis serve action video acquisition image are carried out.The multi-scale detection model of tennis serve action video acquisition image is constructed by using association spectrum mining and detail feature decomposition method.The error recognition of tennis service action video acquisition image is realized.The simulation results show that the accuracy of this method is high and the reliability of error recognition performance is good.