基于模板匹配和边缘检测的车标定位方法  

Approach for vehicle-logo location based on template matching and edge detection

作　　者：李侠[1] 李文举[1] 孙娟红[1] 郑宏亮[1] 

机构地区：[1]辽宁师范大学计算机与信息技术学院,辽宁大连116081

出　　处：《计算机工程与设计》2011年第10期3453-3456,共4页Computer Engineering and Design

基　　金：辽宁省教育厅高等学校科研基金项目(L2010232)

摘　　要：车标定位在车标识别系统中具有关键作用,为解决倾斜的车辆图像及多种车标背景纹理的车标定位问题,提出了一种新的车标定位方法。先进行车辆图像的倾斜校正,再通过三绝对值最大梯度方法定位出车灯带区域,利用车标与车牌、车灯的位置先验知识进行车标粗定位,并根据车标背景纹理特征利用模板匹配和边缘检测方法进行车标精定位。该方法充分利用了车标的先验知识和背景纹理特征,并将模板匹配与边缘检测技术相结合,提高了车标定位的准确率。实验结果表明,该方法既适用于倾斜的车辆图像又适用于多种车标背景纹理,具有很好的车标定位效果。Vehicle-logo Location plays a key rule in the vehicle-logo recognition system. To solve the vehicle-logo location of the tiltvehicle images and multi- types of vehicle-logo background texture effectively, a novel approach of vehicle-logo location is presented according to the vehicle-logo position and the background texture feature. Firstly, the inclined vehicle images are corrected, secondly, the three absolute maximum gradient method is used to find out the car-light area, and then, the priori knowledge about the place relation among vehicle-logos, license plates and car-lights is applied to the rough location ofvehicle-logos, finally, template matching and edge detection are used to locate vehicle-logos precisely according to the background texture of the vehicle-logos. The proposed method takes full advantage of the priori knowledge ofvehicle-logos and combines template matching and edge detection technology, so the accuracy of vehicle-logo location is improved. The experimental results show that proposed method is highly effective for vehicle-logo location because it is fit for either the tilt vehicle images or the multi- kinds of vehicle-logo background texture.

关 键 词：车标定位 倾斜校正 背景纹理 模板匹配 边缘检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]