摘要
提出了基于边缘颜色均值对的车牌定位算法,此方法充分利用车牌的底色与车牌字符颜色的固定搭配,在RGB颜色空间中求取相邻像素的RGB各分量的均值,然后转换到HSV颜色空间,如果出现符合颜色搭配的则认为是车牌可能在的区域.该方法既能保证尽可能多地提取出车牌区域像素,又极大地清除非车牌区域像素,使后续处理得到一定程度的简化.在此基础上利用车牌与车标的位置关系进行车头定位,并结合差分及对称性检测进行车标的精定位.
This paper presents license plate location algorithm mainly based on edge color pair means. This method fully utilizes the collocation of the grounding of the vehicle license plate and the color of the vehicle license plate＇s characters to get the mean value of each RGB component of the neighboring pixels in the color space of RGB, and then transforms it to the color space of HSV If one appears to accord with the collocation of color, which might be regarded as the possible area of the vehicle license plate. This method can draw out the regional pixels of the license plate as many as possible; eliminate greatly the non - vehicle license plate regional pixels, and enable the following processing to obtain some certain degree of simplification. Based on it, to use the relation between the license plate and the vehicle logo locates the position of the locomotive. The vehicle logo location exactly unifies based on the difference and the symmetrical detection.
出处
《曲靖师范学院学报》
2008年第6期68-71,共4页
Journal of Qujing Normal University
关键词
颜色均值对
车牌定位
差分
对称性
车标定位
color pair means
license plate location
difference
symmetrical
vehicle logo location
