基于视觉的大鼠肢体运动自动分析系统  

Automatic analysis system of motion in rat based on vision

作　　者：王松[1] 张恒义[1] 

机构地区：[1]浙江大学生物医学工程教育部重点实验室,浙江杭州310007

出　　处：《传感器与微系统》2011年第11期102-104,107,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(30570433)

摘　　要：设计并实现了实验大鼠肢体运动自动分析系统,系统由通用视频设备和计算机硬件组成,采用基于颜色标记的运动捕捉技术,提出了一种RGB空间映射到HSV空间的快速转换方法,以及快速颜色分类算法结合轮廓搜索算法得到标记点坐标,提取肢体运动学特征参数,并应用于步态行为分类和识别。实际测试平均每帧处理时间小于10 ms,识别正确率97%,表明系统简单、实时和有效。A new method of limb motion automatic analysis system in rat is developed. The system is composed of general video devices and personal computer hardware, based on techniques of motion capture using color marker. The digital image processing such as fast color space conversion,fast color segmentation, and so on, are employed to identify the color markers, extracting features of kinematics parameters are applied in classification and identification of gait phase. The average processing time is less than 10 ms per frame, and the accuracy of recognition is 97 %. Experiment results have shown that this automatic analysis system is simple, real-time and valid.

关 键 词：肢体运动 运动捕捉 行为识别 实时 

分 类 号：TP216[自动化与计算机技术—检测技术与自动化装置]