摘要
按摩机器人在按摩时,由于人体的移动,需要实时对穴位定位和跟踪。本文结合人体背部特征和运动特点,提出了基于模板匹配的按摩机器人穴位定位和跟踪的算法。首先在人体背部预按摩穴位点处布置红色圆形人工标志,利用快速模板匹配算法获取各个人工标志的中心坐标,然后利用相机标定技术获取穴位的世界坐标,最后通过对人体颈部人工标志的跟踪,计算出人体偏移量,将偏移量累加到穴位坐标,实现穴位的跟踪。实验结果表明,该算法具有较高稳定,在误检测方面,明显优于HOUGH算法。
As the movements of the body,it is necessary to get the coordinate acupoint and tracking acupoint when the massage robot is working. An acupoint positioned method based on template matching is proposed,which is combined with human back features and the characteristics of movement. First, making use of fast template matching algorithm to acquire the center coordinates of red artificial mark which put on human back. Second,the world coordinates of acupoint can be get through camera calibration method. Third, camera will track the artificial watchpoint on neck and calculate motion vector which would be added acupoint coordinate. And a cupoint visual tracking can be achieved. Experimental results show that the algorithm has more stability. Template matching is better than HOUGH algorithm in the aspect of error detection.
出处
《科技通报》
北大核心
2011年第5期666-670,共5页
Bulletin of Science and Technology
基金
国家863项目(2010AA040201)
关键词
视觉跟踪
模板匹配
相关度
穴位定位
按摩机器人
visual tracking
template matching
comparability
acupoint positioning
massage robot
作者简介
张化凯（1984-），男，山东鱼台县人，硕士研究生，检测技术、机器视觉研究方向。E-mail：jqrzhk@163．com
