基于模板匹配和边缘检测的车标定位方法.pdfVIP

计算机工程与设计 Computer Engineering and Design 李侠，李文举，孙娟红，等：基于模板匹配和边缘检测的车标定位方法 2011, Vol.32, No.10 3453 基于模板匹配和边缘检测的车标定位方法 + 李 侠， 李文举 ， 孙娟红， 郑宏亮 ( 辽宁师范大学计算机与信息技术学院，辽宁 大连 116081) 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。 关键词：车标定位; 倾斜校正; 背景纹理; 模板匹配; 边缘检测 中图法分类号：TP391.4 文献标识码：A 文章编号：1000-7024 (2011) 10-3453-04 Approach for vehicle-logo location based on template matching and edge detection + LI Xia, LI Wen-ju , SUN Juan-hong, ZHENG Hong-liang (School of Computer and Information Technology, Liaoning Normal University, Dalian 116081, China) Abstract ：Vehicle-logo Location plays a key rule in the vehicle-logo recognition system. To solve the vehicle-logo location of the tilt vehicle 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 of vehicle-logos, finally, template matching and edge detection are used to locate vehicle-