摘要
在车辆牌照识别系统中,由于摄像机畸变、动态范围太窄、车辆牌照被污染等原因,灰度化的车辆牌照图像背景变得模糊,接近于字体的灰度或者动态范围不高,使得前景字体跟背景难以分开。该文采用高帽与低帽形态滤波增强车牌图像中的字体,去除背景对图像的影响,使用基于迭代的图像分块二值化算法进行二值化。实验表明,该算法可有效克服偏暗或泛白背景的影响,二值化效果良好。
In vehicle license plate recognition systems, it is common that the background of the gray level vehicle license plate image becomes unclear or close to the color of the word because of vidicon aberration, narrow dynamic area and polluted plates, which makes it difficult to segment the characters from the plate, therefore a new method is presented, where characters of the image are enhanced by morphologic filtering using the top-hat and the bottom-hat at the same time to eliminate the negative effect of the background, and the plate image is binarized by an improved iterative method. Experimental results show that this algorithm is robust in dealing with the degraded license plates with unclear and white setting.
出处
《计算机工程》
CAS
CSCD
北大核心
2008年第6期210-211,213,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60675024)
关键词
车牌
二值化
局部迭代算法
高帽变换
低帽变换
license plate
binarization
local iterative algorithm
top-hat transformation： bottom-hat transformation
