基于最大梯度差算法的扇形扫描车牌定位方法  

Method of fan-shaped scan license plate location based on MGD algorithm

作　　者：邹晔[1] 宋永献[1] 

机构地区：[1]淮海工学院电子系,江苏连云港222005

出　　处：《计算机工程与设计》2008年第19期5137-5139,共3页Computer Engineering and Design

摘　　要：车牌定位是车牌自动识别系统中关键的第一步,传统的最大梯度差算法按图像像素矩阵的顺序进行扫描和运算,存在运算量大,耗时长的缺点。为此对原算法进行了改进,提出了基于最大梯度差算法的扇形扫描方法,通过改变原算法的扫描方式,减少了对背景噪声的扫描和运算。经实验测试证明该方法运算量小,耗时短,具有较好的应用性。License plate location is the first key step in the automatic license plate recognition system. The traditional MGD （maximum gradient difference） algorithm based on the order of the image pixels matrix has the shortcomings of the large operation measure and long- time running. A new improved fan-shaped scan method is presented. By changing the scan method of the original algorithm, the scan and operation to the background noises are greatly decreased. Testing has approved the validity of this method-- smaller operation measure and shorter running time.

关 键 词：车牌定位 图像 最大梯度差 闽值 扇形扫描 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]