一种改进的车牌图像二值化及定位算法  

An Improved Algorithm of Vehicle License Plate Location

作　　者：鹿昌义[1] 李敏[1] 黄廷磊[1] 

机构地区：[1]桂林电子工业学院计算机系,广西桂林541004

出　　处：《桂林电子工业学院学报》2004年第6期36-39,共4页Journal of Guilin Institute of Electronic Technology

摘　　要：就车牌图像中的光照和污损的复杂特征,对车牌预处理、边缘提取及二值化和车牌定位进行了一系列分析研究。通过设计一种边缘提取与二值化结合的预处理算法,利用提出改进的增加三角点的角点定位算法,设定优先级的四角点、三角点和对角点方法对车牌进行定位。经改进后的算法,能有效对车牌鉴别算法对搜索到的车牌区域进行鉴别,从而提高了车牌定位的准确率和速度。This paper analyzes the problem of image preprocessing, edge extraction and binarization, as well as vehicle license plate location with complicated-background image. Then this paper presents a set of algorithm. Firstly, a kind of algorithm of image preprocessing of edge extraction with binarization is designed; secondly, an improved algorithm of vehicle license plate location by extracting four corner-points, three corner-points, two corner-points is designed and at the same time an algorithm of vehicle license plate discrimination is designed. The result shows that this algorithm can effectively approve the accuracy and speed of vehicle license plate location.

关 键 词：角点提取 二值化 车牌定位并提出 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]