摘要
为了能在复杂背景和不同光照条件下快速、准确地定位车牌位置,提出一种基于行扫描和小波变换的车牌定位算法。该方法首先对车牌图像实行边缘检测、二值化等预处理,在此基础上,利用行扫描法确定车牌的上下边界,对水平区域图像的垂直投影进行小波变换,设定阈值,确定左右边界。实验结果表明,该方法定位速度快、准确率可达到98%,具有很好的鲁棒性。
In order to locate license plates rapidly and accurately in complex background and variable conditions of lights,a novel car license plate localisation algorithm based on line scanning and wavelet transform is introduced.In this algorithm,some pre-processing including edge detection and binary image restoration on the image of the license plate is done firstly.Then,up and down boundaries of license plate are decided based on the line scanning method.At the same time,left and right boundaries of license plate can be figured out by wavelet transform conducted on vertical projection of the horizontal area image as well as by setting the threshold.Simulation results show that this plate localisation algorithm has fast localisation speed,high accuracy up to 98% and robustness.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第5期279-282,共4页
Computer Applications and Software
关键词
车牌定位
行扫描
投影
小波变换
Car plate localisation Line scanning method Projection Wavelet transform
