摘要
基于双目立体视觉的障碍物检测方法,是一种基于双目视觉的V视差图检测障碍物的算法。利用生成的视差图之后再生成V-视差图,进而提取V-视差图当中的直线的信息,利用这些提取的直线的信息,就可以大致确定障碍物存在的区域,以此进行障碍物的检测。这样是一种对于光照和阴影等的干扰不是十分敏感,可以检测具备面特征的障碍物,可以用于复杂的背景之下的障碍物的检测。
The obstacle detection method based on binocular stereo vision is an algorithm for obstacle detection based on binocular vision V-parallax map.The V-disparity map is generated using the generated disparity map,and further,the information of the straight line in the V-disparity map is extracted,and by using the extracted information of the straight line,the region in which the obstacle exists can be roughly identified,thereby making the obstacle detection.Such a light and shadow interference is not very sensitive,can be detected with surface features of the obstacles,can be used for the detection of obstacles under the complex backgr ound.
作者
邓博
吴斌
Deng Bo;Wu Bin(Southwest University of Science and Technology,Mianyang Sichuan 621010,China)
出处
《信息与电脑》
2018年第1期41-42,45,共3页
Information & Computer
关键词
障碍物检测
双目视觉
立体视觉
V-视差
obstacle detection
binocular vision
stereoscopic vision
V-parallax
