基于双目视觉和Voronoi图的移动机器人地图构建研究  

Study on Mobile Robot Mapping Based on Binocular Vision and Voronoi Diagram

作　　者：于乃功[1] 李博[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《计算机测量与控制》2012年第10期2766-2768,共3页Computer Measurement &Control

基　　金：北京市教育委员会科技发展计划面上项目(KM200810005016)

摘　　要：在双目视觉传感器的基础上,应用改进后的SIFT(Scale-invariant feature transform)算法确定障碍物的三维坐标,提出最大包围盒的思想,将处理后的数据作为Voronoi图的建模生成元,完成移动机器人的全局地图构建;基于V图构建的地图的最大的优点是移动机器人行走的路径是障碍物离散中心的中垂线,可以保证机器人在运行过程中最大程度地远离障碍物;试验结果表明应用该方法构建的地图可靠性高,符合导航的要求。Determining the three--dimensional coordinates of the obstacle with binocular vision sensor based on the application of im proved SIFT （Scale--invariant feature transform） algorithm, the data will be treated as the the model generator of Voronoi diagram to corn plete the glohal map building. The biggest advantage of the V diagram is that the environmental path road is based on the vertical of the dis crete centers of obstacles, which may make sure that the robot can keep away from the obstructions during operation. The experimental re suit shows that the map after construction demonstrate high reliability which consistent with the requirements of navigation.

关 键 词：双目立体视觉 SIFT算法 VORONOI图 地图构建 

分 类 号：TH873.7[机械工程—仪器科学与技术]