中文摘要：

基于视觉系统的自主导航技术是近年来机器人研究领域的一个热点．以双目视觉为基础，采用张正友棋盘标定法实现了立体摄像机的精确标定，基于区域的归一化互相关法完成视频的立体匹配，并通过盒滤波技术改善了算法的执行速度，基于膨胀栅格的距离向量直方图法实现运动系统最优路径的选择．通过使用标准测试图和真实场景图对系统性能进行测试，结果表明算法符合要求．最后给出了系统的实现方案．

英文摘要：

The robot autonomous navigation technology based on visual system has become a hot research field in recent years. Based on binocular stereo vision, chessboard calibration method by Zhang Zhengyou is used to achieve precise calibration of a stereo camera. The area-based normal- ized cross-correlation method is utilized to realize the video stereo matching. In order to improve the execution speed of the algorithm, the cartridge filter technology is used. The distance vector histo- gram method based on expansion raster is utilized to achieve optimal path selection motion system. The performance of the system is tested by comparing the standard graph with the real scene. And the results show that the system can get good autonomous navigation movement.