摘要
以足球机器人视觉系统为研究对象,提出了一套基于颜色信息的色标识别方法。在假设颜色样本图像中存在唯一主色的条件下,提出了排异聚类主色提取算法,该算法适合足球机器人视觉系统的色标颜色提取过程,具有抑制噪声和保持主要颜色信息的能力。基于颜色相似系数判定图像中待识别像素颜色与样本颜色之间的相似性。采用网格搜索方法实现色标的搜索过程。通过形态学滤波方法对色标预分割结果进行后处理。实验结果表明,该方法具有较理想的彩色色标识别效果。
With the soccer robot vision system as studied object, a set of methods of recognition of color mark is proposed. On assumption that there is a uniquely main color in color pattern image, an algorithm of excludible clustering main color extraction is brought forward to fulfill the color sampling for color mark of soccer robot. The algorithm is suited for the extraction of pattern color of color object in vision system of soccer robot, and has capability of restraining noise and maintaining main color information. Based on color similarity coefficient, the similarity between two kinds of color of a pixel in a color image and the pattern extracted is estimated. Method for grid searching is adopted to perform color mark searching. The rough segmentation result is furthermore processed by morphology filtering technology. Experimental results show that the approach presented has ideal precision for color mark recognition.
出处
《工程图学学报》
CSCD
北大核心
2005年第3期100-105,共6页
Journal of Engineering Graphics
关键词
计算机应用
色标识别
彩色图像分割
足球机器人
computer application
recognition of color mark
color image segmentation
soccer robot
