一种基于色差灰度图像的足球机器人足球识别方法  

A Football Recognition Method for Soccer Robot Based on Color Aberration Gray-Level Image

作　　者：刘波[1] 胡爱萍[2] 范骏波[1] 

机构地区：[1]河海大学机电工程学院,江苏常州213022 [2]江苏工业学院,江苏常州213016

出　　处：《江苏工业学院学报》2009年第4期68-71,共4页Journal of Jiangsu Polytechnic University

摘　　要：针对足球机器人足球识别问题,提出构建R-G色差灰度图像,通过灰度阈值分割和消除小面域,较好地实现对Robo-Cup中型组规定比赛用球的快速准确识别。同时,针对比赛环境中场外与足球近似颜色物体的干扰,结合全景图像特征,采用目标面域大小和图像坐标系下的像素距离构建阈值滤波,较好地解决了场外物体对目标足球识别的干扰问题。Aimed to solve football recognition of soccer robot, a novel method was proposed. Firstly a color aberration gray--level image was set by subtracting G components from R components. Then, gray-- level threshold filtering by setting proper threshold value and blob filtering were processed. Recognition of football specifying in RoboCup MSL competition was realized accurately and quickly. Furthermore, considering the characteristic of omni--vision image, a threshold filter method was also proposed, which con- structed threshold using the size of objective region and pixel distance between the center of objective re- gion and the center of the image under image coordinate. By using this method, the disturbance caused by objects outside the ground whose color was similar to that of the ball in the competition circumstance was filtered in most cases.

关 键 词：足球机器人 足球识别 RG色差图像 全景图像 干扰滤除 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]