基于最大梯度差和边缘搜索的车牌定位方法  

Vehicle license plate location algorithm based on maximum gradient difference and edge search

作　　者：甘玲[1] 林小晶[1] 

机构地区：[1]重庆邮电大学计算机科学与技术学院,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2010年第3期371-374,共4页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金项目资助(60873186)~~

摘　　要：车牌定位是车牌识别系统的关键技术,定位的准确与否直接影响车牌识别的结果。为了实现车牌的准确定位及车牌识别,依据车牌区域具有丰富的边缘信息,提出了一种改进的车牌定位算法。利用最大梯度差缩小车牌图像范围和矩形匹配法粗定位车牌,根据搜索车牌的边缘信息来精确定位车牌。实验结果证明了该方法的可行性及定位的准确性,并且比单一使用最大梯度差法的定位准确率有较大的提高。The vehicle license plate location is the key technique of the vehicle license plate recognition system.The accuracy influences the plate recognition result.In order to realize license plate location exactly,an advanced license plate location method was proposed according to the edges information in the area of license plate.First,shrink the searching area of license plate with the maximum gradient difference.Second,gain the approximate area of license plate with the rectangle matching method.Finally,position plate accuracy by searching the edge information of license plate.Experiment results indicate that the algorithm is feasible,more accurate than that only with the method of maximum gradient difference.

关 键 词：车牌定位 最大梯度差 矩形匹配 边缘搜索 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]