摘要
近年来,关于自主机器人动态目标跟踪方面的研究越来越多,但是动态目标跟踪的实时性和准确性仍然是亟待解决的问题。研究了基于激光扫描的动态目标跟踪方法,提出了类圆柱体人腿检测法,给出了该方法的详细过程,具体包括利用级比与阈值相比较的方法分割出被检测物体的边界,基于空域聚类法的类圆柱体人腿检测以及基于时域聚类法的人腿跟踪。最后,通过实验验证了基于激光扫描的动态目标跟踪方法的有效性,使机器人在实现动态目标跟踪时,能够适应更多复杂、变化的室内外环境。
In recent years,there are more and more research on autonomous robot dynamic target tracking.But the real-time per- formance and the accuracy of the dynamic target tracking is still a difficult problem to be solved.A dynamic target tracking methods based on laser scanning has been developed.It proposed the cylinder-like detection method of a human leg,and give the detailed process of the method,including the use of compare stepwise ratio with threshold method which success- fully splits the boundary of the object to be detected,cylinder-like legs detection based on spatial domain clustering method,legs tracking based on time domain clustering method.Finally,some experiments have been made to verify the ef- fectiveness of the method for target tracking based on laser,which shows these methods can make the robot adaptable to the more complicated,changing indoor and outdoor environment.
出处
《工业控制计算机》
2014年第1期79-80,82,共3页
Industrial Control Computer
关键词
自主机器人
目标跟踪
激光扫描
autonomous robots,target tracking,laser scanning
