车牌图像的局部倾斜畸变校正方法  

Local correction approach for inclined and distortional VLP image

作　　者：李元金[1,2] LI Yuanjin 

机构地区：[1]滁州学院计算机科学与技术系,安徽滁州239000 [2]东南大学影像科学技术实验室,南京210096

出　　处：《计算机工程与应用》2011年第17期10-12,17,共4页Computer Engineering and Applications

基　　金：安徽省教育厅自然科学研究项目(No.KJ2009B107Z);应用数学省级本科教学团队(滁州学院)

摘　　要：针对车牌图像的倾斜畸变会对后继字符切割和识别结果带来不利影响的问题,提出使用局部技术和适度缩放技术对车牌图像倾斜畸变进行校正。根据车牌图像不同位置倾斜畸变程度不一样的特点将车牌图像分成若干份;分别对每份图像使用"连接点"法进行倾斜畸变校正,根据字符高度的变化对字符宽度进行同样的缩放;将校正之后的各个子图像连接在一起成为校正之后的车牌图像。实验表明,使用局部法对倾斜畸变车牌图像进行校正的效果要优于使用全局法。In the light of the problem of inclined and distortional Vehicle License Plate（VLP） image that brings adverse effect on subsequence steps of characters segmentation and recognition, this paper proposes an approach based on the local and adaptive scaling techniques to correct inclined Vehicle License Plate（VLP） image.According to the inclined and distortion condition,the method,firstly,divides the VLP image into several parts,and then uses connection point techniques to correct every sub-image and adaptively resizes for the width of every sub-image based on the change of its height.Finally, the corrected sub-images are linked into a whole image.Experiment showes that,using the local technique to correct the inclined and distortion VLP image can obtain better effectiveness than global method.

关 键 词：适度缩放技术 倾斜畸变校正 局部技术 全局法 车牌图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]