基于小波变换和亮度矩的车牌图像分割算法  

AN ALGORITHM OF SEGMENTATION OF VEHICLE-LICENSE-PLATE IMAGES BASED ON WAVELET TRANSFORM AND INTENSITY MOMENT

作　　者：崔新春[1] 魏春英[2] 

机构地区：[1]曲阜师范大学信息技术与传播学院信息技术系,曲阜市273165 [2]济宁师范专科学校计算机中心,山东省济宁市272025

出　　处：《曲阜师范大学学报（自然科学版）》2002年第4期75-78,共4页Journal of Qufu Normal University(Natural Science)

摘　　要：车辆牌照图像的分割是车牌识别的前提和基础 .针对牌照与车身背景的分割问题 ,该文提出了基于小波变换和亮度矩方法的车辆牌照阈值分割算法 .首先 ,对车牌图像进行小波去噪 ,然后利用亮度矩对去噪后的车牌图像进行阈值分割 .实验结果表明 :此方法能够从含有较强噪声的车辆图像中获取车牌图像 ,且具有相对于图像的平移。With the development of the technology of automatic management of vehicles, more and more attention are placed on segmentation of vehicle_license_plate images as it is the basis of vehicle_license_plate images recognition. In this paper, an algorithm of segmentation of vehicle_license_plate images based on wavelet transform and intensity moment is presented. First, apply wavelet transform on a vehicle_license_plate image to denoise it, then threshold the image on condition of keeping the intensity moment invariable. Experiments show that the method can get vehicle_license_plate image out of rough vehicle image even when it was badly noised, and this method has low sensity to shift, rotation and scale change.

关 键 词：车牌图像 车牌识别 小波变换 亮度矩 阈值化 图像分割 阈值分割算法 小波去噪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]