货车图像车牌区快速定位及字符切割算法  

Algorithm for fast location of freight license and character segmentation

作　　者：付仲良[1] 黄书强[2] 陈珉[2] 陈江平[1] 

机构地区：[1]武汉大学空间信息与数字工程研究中心,湖北武汉430079 [2]武汉大学计算机学院,湖北武汉430079

出　　处：《计算机工程与设计》2003年第1期77-79,共3页Computer Engineering and Design

摘　　要：针对货车图像中的光照不均匀和各种污损的条件,重点讨论了货车图像的二值化、车牌区定位及字符切割方法,并提出了一种货车车牌区定位快速定位算法(TLFP)和一种投影与弧特征相结合的字符切割算法。大量实验表明:使用上述算法,车牌区定位准确率达99.7％,字符切割正确率达98.5％。Aiming at the situation of the lighting asymmetry and the variety of dirt in the truck image, the paper mainly discusses the methods about image binarization, license plate region location (LPR) and character segmentation. On the basis of the discussion, this paper presents an algorithm of fast location of license plate region and an algorithm of character segmentation based on projection and feature of arc. Large quantity of experiments have indicated that by using the above algorithms the accuracy rate of the location of plate region reaches 99.7% and the accuracy rate of character separation may reach 98.5%.

关 键 词：货车图像 车牌区 快速定位 字符切割算法 铁路运输 运输管理自动化 

分 类 号：U29－39[交通运输工程—交通运输规划与管理]