一种基于数学形态学的车牌图像分割方法  

Segmentation Method of Vehicle-License-Plate Images Based on Mathematical Morphology

作　　者：王钧铭[1] 赵力[2] 

机构地区：[1]南京信息职业技术学院,江苏南京210013 [2]东南大学,江苏南京210096

出　　处：《电视技术》2007年第10期84-86,共3页Video Engineering

基　　金：教育部博士点基金资助(20050286001)

摘　　要：提出一种基于数学形态学的车牌图像分割提取方法。用修正后的形态学边缘检测算子对图像进行边缘检测,采用二尺度结构元素检测平均方法提高边缘检测的准确性,再用不同的结构元素对边缘图像进行形态滤波,以消除干扰。实验证明,该方法能快速准确定位分割出车牌图像,且计算量较小。The proposed segmentation method of Vehicle-License-Plate images is based on mathematical morphology. It first detects the edges of image through an amended morphology arithmetic operator. And then, an average method of two structure elements detecting with different scales is used to improve the veracity of edge detecting. Finally, the processed edge image is morphologically filtered by different structure elements to eliminate interference. During the experiment, the method can separate the image of license plate quickly, accurately and with less calculation.

关 键 词：车牌分割 数学形态学 边缘检测 结构元素 

分 类 号：TN911.73[电子电信—通信与信息系统]