自主足球机器人视觉系统结构及关键技术

提出了自主足球机器人嵌入式多处理器体系结构以及视觉系统的构成框架.对单目视觉系统所涉及的关键技术进行了详细的分析和研究,其中包括图像分割、目标搜索算法、目标定位和机器人自定位技术.基于YUV颜色空间,对目标进行分割,并通过连通区合并来快速搜索目标,同时根据摄像机模型来确定目标的空间位置.实验表明分布式视觉体现结构能够实时准确地完成视觉处理任务.以上新算法可以快速鲁棒地搜索图像目标,并能较精确地确定目标的位置信息和机器人的自定位信息.

Framework and Key Technologies of Autonomous Soccer-Robot Vision System

The system structure of a multi-processor embedded autonomous soccer-robot and relevant framework of vision system are proposed. The key technologies of single camera vision system are analyzed in detail, including the color image segmentation, algorithms for searching objects, object localization and robot self-localization. The objects are segmented on the basis of YUV color space. Moreover, the objects are searched by way of merging the connexity regions. The linearity model of camera is utilized to locate the spatial position of objects. The experiment results showed that the image objects can be found accurately real-time and that the newly developed algorithm can robustly and quickly search image objects and acquire the comparatively accurate information on their positions and robot self-localization.