白点定位图像处理算法  

Image Processing Algorithm of White Point Localization

作　　者：吴晓[1,2] 曹其新[1] 

机构地区：[1]上海交通大学机械与动力工程学院,上海200240 [2]蒲田学院,福建蒲田321100

出　　处：《中国矿业大学学报》2008年第6期830-833,847,共5页Journal of China University of Mining & Technology

基　　金：国家自然科学基金项目(60443007)

摘　　要：为了解决全自主足球机器人自定位难题,采用白点定位图像处理算法求解足球场地白线边缘中点作为自定位特征点.通过沿预定义一维扫描线的彩色滤波、求最大梯度点和亚像素分解求中点等3个步骤,求解出若干定位白点,为机器人自定位做准备.结果表明:白点定位图像处理算法,定位精度最大误差为1.2mm,程序运行平均时间为15ms,定位过程抗干扰能力强,完全能够满足足球机器人自定位准确性、实时性和鲁棒性的要求.In order to solve the problem of self localization for autonomous robot soccer, an image processing algorithm of white point localization was applied to find out the midpoint of white line edge of the soccer field as the reference point for self localization. By following colored filter of predefined single-dimension scanline, calculating the maximum gradient point,and deriving the midpoint through sub-pixel decomposition, several localizing white points were derived to prepare for the robot＇s self-localization. The results show that the image processing algorithm of white point localization has a maximum localization accuracy error of 1.2 ram, an average program-running time of 15 ms, and is strongly resistant to interference in the process of localization. The image processing algorithm of white points localization can completely satisfy the requirements of accuracy, realtime and robustness for robot soccer in self-localization.

关 键 词：白点定位 图像处理 亚像素 机器人 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置] TS251.52[自动化与计算机技术—控制科学与工程]