一种基于自适应形态学结构元素的车牌定位算法  

License plate location algorithm based on adaptive structure element of mathematical morphology

作　　者：方兴林[1] 

机构地区：[1]黄山学院经济管理学院,安徽黄山245041

出　　处：《计算机工程与应用》2013年第23期149-152,共4页Computer Engineering and Applications

基　　金：黄山学院校级科学研究项目(No.2012xkj004)

摘　　要：针对传统的基于数学形态学的车牌定位算法结构元素选取的弊端,即依据经验或实验而选择一个固定大小和形状的结构元素对车牌图像进行全局处理,导致定位算法的普适性和鲁棒性低,提出了一种基于自适应结构元素的车牌定位算法,算法充分利用车牌二值图像自身信息,通过计算字符水平边缘所形成的线段长度的均值作为结构元素,因此结构元素的大小能随着车牌图像的不同而自适应调整,更加有效地定位车牌目标区域。实验结果证明,算法具有较强的自适应能力,定位准确率满足应用需要。In view of disadvantage of selecting structure element of the traditional license plate locating algorithm based on mathematical morphology, that is, it selectes a definite shape and fixes size structure element based on experience or experiment results to globally process the license plate image, and the disadvantage would reduce the robustness and universal, a license plate location algorithm based on adaptive structure element is proposed. The algorithm makes the best of self-information of binary license plate image. The mean values of lines length, which are formed by calculating horizontal edge of characters, are as the structure elements, so the size of each structure element is adaptive to different images, in order to effectively extract the true license plate regions. The experimental results show that the algorithm has stronger adaptive ability, and the accurate rate of localization meets application requirements.

关 键 词：结构元素 车牌定位 自适应算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]