基于全向视觉信息的足球机器人自定位方法研究  

Research of Self-localization System Based on Omnidirectional Vision for Soccer Robot

作　　者：蔡宗琰[1] 马鹏飞[1] 欧阳敏[2] 邓小康[1] 

机构地区：[1]长安大学道路施工技术与装备教育部重点实验室,西安710064 [2]神龙汽车有限公司技术中心,武汉430000

出　　处：《科学技术与工程》2015年第29期57-61,66,共6页Science Technology and Engineering

基　　金：国家级大学生创新创业训练计划项目(201410710020)资助

摘　　要：RoboCup中型组比赛对抗激烈。针对赛场的高度动态性,提出了一种快速的机器人自定位方法。首先根据全向视觉机器人的结构特点。在进行视觉处理时,利用图像掩模去除无效信息;同时针对全像素搜索的不必要性,采用基于Bresenham算法的像素搜索方法,缩小像素搜索范围。在机器人自定位时,利用白线定位方法,结合基于Bresenham算法的像素搜索方法,有效的提高图像处理的实时性。最后,以ASROⅡ机器人为实验平台,设计了实验验证了该方法不但提高了实时性;而且完全可以满足比赛对定位精度的要求。In order to solve the self-localization problem in highly dynamic situation in the Robo Cup middle size league competition,a fast robot self-localization method was used. Firstly,according to the structural of omnidirectional visual robot during visual processing,using an image mask the invalid information was removed. In order to avoid searching for all the pixels and narrow the scope of pixel,a method based on Bresenham algorithm was used.Using white line positioning algorithm to determine the robot position,combined with the search method based on Bresenham algorithm,the real-time image processing is effectively improved. Finally,a test was design based on ASROⅡ robot and it is proved reliable and effective.

关 键 词：全向视觉信息机器人 BRESENHAM算法 扫描线 白线检测 自定位 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]