摘要
本文针对车牌图像的特点提出了一种针对字符分割的车牌图像预处理方法。该方法首先将定位出的彩色图片进行彩色图像的灰度化处理,再进行目标增强、二值化、去除铆钉干扰、倾斜校正,从而实现车牌的精确定位,为字符分割做好准备。实验结果表明,该方法处理效果好,对光照不均、对比度低、污迹及倾斜度较大的车牌图像具有很好的鲁棒性。
The article aims at the characteristics of license plate images,and puts forward a new approach to license plate preprocessing by character segmentation. The license plate after location is usually colorful, so before the segmentation isdone, we transfer the color plate to gray firstly, and then enhance the image, threshold, clear the rivet, tilt correction and obtain a refined location of the license plate. Experimental results show the proposed method is easy-to-implement and offers robustness when dealing with the number plate that has dirt, poor contrast, uneven illumination or tilt.
出处
《计算机工程与科学》
CSCD
2008年第4期40-42,共3页
Computer Engineering & Science
关键词
车牌字符分割
图像增强
二值化
倾斜校正
license plate character segmentation
image enhancement
threshold
tilt correction
