为了实现Pioneer3-AT轮式移动机器人避障的目的,设计了一种移动机器人避障的系统方案.该方案采用一种快速融合图像分割和立体视觉的障碍物定位方法,即利用经典图像处理把障碍物从背景中分割出来,同时提取其轮廓信息进行立体匹配,之后结合摄像头标定结果,实现空间点的三维重建.采用双目视觉和声纳传感器相配合来获得周围障碍物的信息,最终根据模糊控制原理设计了避障规则和避障控制器.通过实际的躲避障碍物实验,实现了移动机器人成功躲避障碍物的功能.实验结果证实了该系统的可行性和有效性.%A method of obstacle positioning based on fast fusion of image segmentation and stereo vision is put forward in this paper to realize obstacle avoidance of Pioneer3-AT wheeled mobile robots, namely separating obstacles from the background by typical image processing method, and performing the stereo matching of the extracted contour features in combination with the binocular vision calibration results to get the spatial point of 3D reconstruction. The binocular vision and sonar sensors are adopted in the obstacle avoidance technology to obtain the information about the obstacles in the vicinity of the robot at the same time. The obstacle-avoidance rule and controller according to the fuzzy control principle were designed to avoid collision and bypass the enroute obstacles. The function of evading obstacles was successfully achieved for the moving robots. The experiment shows the feasibility and validity of this system.
展开▼
