基于图像的六足昆虫的步态检测与三维重构  

A Method Based on Image Process of Hexapod Insect in Gait Detection and 3D Reconstruction

作　　者：陆汉雄[1] 何斌[1] 黎明和[1] 

机构地区：[1]同济大学电子与信息工程学院,上海201804

出　　处：《计算机与数字工程》2010年第2期114-117,共4页Computer & Digital Engineering

基　　金：国家自然科学基金"新型湿吸机器人驱动原理研究"(编号:50405045);国家"八六三"计划"仿生湿吸微型机器人"(编号:2007AA04Z253)资助

摘　　要：采用图像方法对六足昆虫进行步态规划,分析图像的颜色空间,选择标记点的颜色,根据颜色向量在颜色空间的夹角,提取关节点。利用区域搜索算法,检测轮廓,计算区域的质心,采用余弦定理计算各质心的夹角,结合俯视图、侧视图实现三维重构。实验表明该方法简单有效,最大误差10%,满足仿生机器人的步态规划要求。The gaits of hexapod insect are planned by image process. The color space of image is analyzed and a proper color for marked points is chosen. According to angles between vectors, the joints are extracted. The centroids of contours are calculated by region search algorithn~ And the angles between centroids are got. The 3D reconstruction is realized with vertical and side views. The results of experiment show that the method is simple and effective. Maximum error is below 10%. It meets the requirements of bio-robot gait planning.

关 键 词：颜色空间 运动检测 区域搜索 三维重构 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]