基于多尺度小波边缘检测的车牌定位研究  

Research on Vehicle License Plate Location Based on Multi-scale Wavelet Edge Detection

作　　者：张吉斌[1] 滕青芳[1] 王俊臣[1] 张佩炯[1] 

机构地区：[1]兰州交通大学自动化与电气工程学院,甘肃兰州730070

出　　处：《电视技术》2013年第7期150-153,共4页Video Engineering

摘　　要：车辆牌照的准确定位是车牌识别系统中的关键步骤,利用车牌区域丰富的边缘和纹理信息以及车牌自身的特征,提出一种基于多尺度小波边缘检测的车牌定位方法。该方法能够更好地解决在复杂背景和复杂光照下的车牌定位。首先用图像增强和多尺度小波算子提取出车牌图像的边缘,然后利用数学形态学和连通区域标记的方法对车牌进行初步特征提取去除伪车牌区域,最后采用水平垂直投影法进行车牌的精确定位。实验结果表明,该方法能够实现车牌的快速准确定位,对复杂背景下的车牌具有很好的鲁棒性和实时性。Vehicle license plate of accurate location takes an important part in the vehicle license plate recognition system, using the license plate area rich edge and texture information and license plate own characteristic, a method of vehicle license plate location is put forward based on multi-scale wavelet edge detection. This method can better solve license plate location in the complex background and complex light. Firstly, license plate image edge is extracted through the image intensification and multi-scale wavelet operator, and uses the mathematical morphology and connected components la- beling method for preliminary characteristics vehicle license plate image to remove pseudo-plate region, and then uses horizontal and vertical projection method to separate the accurate license plate. The experiments show that the method proposed can realize the rapid and exact license plate location, the algorithm has strong robustness and real-time for complex background of license plate.

关 键 词：车牌定位 多尺度小波 数学形态学 边缘检测 

分 类 号：TN911.73[电子电信—通信与信息系统]