【摘要】 为了在不干扰正常训练的情况下,实现数字铅球系统中的"力-姿"信息同步,提出了一种基于图像特征的非侵入式同步方法。采集高速图像序列后,利用圆度特征判决铅球出手帧,通过ROI设置及关键帧预判等方法提高算法效率,实验中对比分析了参数设置对算法效率的影响。实验结果表明,算法实现了高速高清图像的实时处理,同步准确率达到93.1%,有效地满足了辅助训练的需要。更多还原

【Abstract】 To solve the"force-posture"synchronization problem in the digital shot system without disturbing the normal training,a non-invasive data-image synchronization method based on image features is proposed.The key frame of shot-putting is judged by the roundness feature after the high-speed image sequences captured,and the approaches such as ROI setting and key frame predicting are employed to improve the algorithm efficiency for real-time high-speed and high-resolution processing.Then the influence of parameters setting on algorithmic efficiency is analyzed comparatively in experiments.The experimental results show the synchronization accuracy can reach 93.1%,which meet the needs of auxiliary training effectively.更多还原