摘要
对机器人视觉导航而言,道路识别和表示是一个非常重要的环节,它直接影响到后续的路径规划。该文针对红外道路图像,提出了基于区域方法的一套处理方案,该方法首先通过分割获得道路区域,利用链码跟踪获取道路边缘的链码。采用了一种通用的道路模型,然后基于链码以及该道路模型,设计了一种有效的道路边界拟合方法。在拟合过程中,首先依据一定的准则把链码分为两段,对于每一段再递归执行该分段过程,直到不能分为止,然后用分段直线去描述道路边界。该拟合算法可以有效地处理直道和非直道的情况。文中给出了相关的实验结果。
For robot vision navagation, the recognition and representation of road is a very important step, which influences the subsequent path planning directly. A set of effective processing methods are proposed for infrared road image, which is a region-based method instead of edge-based one. Segmentation is executed to obtain the road region and then road boundary is traced by chain code. One general road model is adopted, and based on the chain code and this road model, one fitting algorithm of road boundary is designed. Firstly, the fitting algorithm divides the chain code into two parts by certain criterion, such procedure is applied recursively to each part until it can＇t be divided further, then line segments are used to describe the road boundary. This fitting method can accomodate both straight line and non-straight line effectively. The related experimental results are presented too.
出处
《计算机工程与设计》
CSCD
北大核心
2006年第15期2812-2815,2831,共5页
Computer Engineering and Design
关键词
红外图像
道路识别
道路模型
拟合
视觉导航
infrared image
, road recognition
road model
fitting
vision navigation
