摘要
车牌定位是汽车牌照识别系统的关键技术之一。通过对图像进行预处理,去除干扰的噪声,然后利用隔行扫描和矩形窗口搜索法来定位车牌的上下边界和左右边界,从而提高了车牌定位的速度。提出了利用车牌定位出来的区域面积的变化情况来判断车牌的倾斜情况,并对倾斜的车牌进行了准确的校正。实验结果表明,该算法简单高效。
License plate location is one of the key technologies in license plate recognition system. By preprocessing the license plate image to remove the interference noise, this paper uses the method of interlaced scanning and the method of rectangular window searching to location the license plate＇s top boder, bottom boder, left boder and right boder. So it greatly improves the speed of license plate location. This paper also proposes that the changing of location area can judge the tilt condition of the license plate and gives out a correction method of the tilt license plate. The experiments show that this algorithm is simple and efficient.
出处
《计算机系统应用》
2011年第9期214-217,233,共5页
Computer Systems & Applications
关键词
车牌定位
二值化
隔行扫描
矩形窗口
倾斜校正
license plate location
binarization
interlaced scanning
rectangular window
tilt correction
