基于标志点识别的自主车视觉导航

提出了一种基于室内标志点识别的单目视觉导航的方法。在地面上有规律地旋转一些标志点,采用辅助先验知识对标志点进行识别和分割,利用透视投影变换得到标志点与摄像机之间的相对距离,达到通过单目摄像机获得景物深度信息的目的。整个算法结构简单,易于实现,对由照明不均匀和地面反光所引起的各种干扰有较好地抑制作用。

Visual navigation of autonomous mobile robot based on marking-points recognition

A monocular visual navigation of autonomous mobile robot based on recognizing marking-points indoors is presented.Arranging marking-points on the floor in a rule,the images was gathered by monocular camera.It was important to combine the feature prior knowledge with image segmentation methods to process the gathered images and recognized the marking-points.The relatively distance between camera and marking-points was gained by removing the perspective projection effects,thus the depth of scenery was determined.Experimental results show that the algorithm is simple and easy to realize.It is available for suppression of the influences from the uneven lights and noises.