基于全向视觉的移动机器人实时全局地图构建

Global Map Building Based on Omni-Vision for Mobile Robot

实时全局地图构建是实现移动机器人智能化的关键;研究了一种基于全向视觉的移动机器人全局地图的实时构建方法.首先介绍了全向视觉的体系结构,然后对图像处理软件的相关模块,包括基于颜色的阈值分割、区域连通、特征提取和目标识别等进行了说明,最后通过坐标的转换实现全局地图的构建;实验结果表明,由于充分利用了全向视觉的特性,构建的全局地图准确性高、实时性好,完全能够满足移动机器人对环境建模的需要。

Global map building in real time is essential task for an autonomous mobile robot, machine vision is the symbol of intelligent robot. This paper introduces the omni--vision system firstly, and then explains the ways of calculating every module including thresholding segmentation based on color, region growing, character extraction and object recognition in detail, and mapping building are realized through Coordinates transformation at last. Experiment results show that global map building has good accuracy and real time for the best u sing of omni--vision.