【作者】 董华； 王人成； 黄昌华； 金德闻；

【Author】 Dong Hua, Wang Ren Cheng, et al.Department of Precision Instruments and Mechanology, Tsinghua University,Beijing 100084

【机构】 清华大学精密仪器与机械学系；

【摘要】 目的 研究一种快速识别人体运动图像标志点的方法,并通过实验分析探讨标志点的大小、颜色、形状、变形等对识别结果的影响,进而提出对实验光线及背景的具体要求.方法 在进行图象预处理后,运用广义Hough变换检测标志点边缘.结果 平均每帧图像识别时间小于1秒,识别率超过95%.结论 采用广义Hough变换进行标志点识别是一种快速有效的方法.更多还原

【Abstract】 Objective Our goal is to develop a fast method of recogniing markers from human motion, to analyze the fundamental characteristics of markers such as size, color, shape and deformation etc. and to educe the specific requirements of the experimental illumination and background. Method After image preprocessing, the Generalized Hough Transform technique is used to detect the edges of markers. Result The experimental result shows that the average recognition time of the images is within 1 second and the recognition rate is greater than 95%. Conclusion The method of using Generalized Hough Transform to recognize markers is validated.更多还原