基于颜色特征和改进Canny算子的车牌图像定位  

Recognition of Vehicle License Plate Locating Based on Color Feature and Improved Canny Operator

作　　者：孙金岭[1] 庞娟[1] 张泽龙[1] 

机构地区：[1]兰州理工大学经济管理学院,兰州730050

出　　处：《吉林大学学报（理学版）》2015年第4期693-697,共5页Journal of Jilin University:Science Edition

基　　金：甘肃省高等学校科研基金(批准号:2014B-037);甘肃省社科基金(批准号:14YB050);兰州市社科规划项目(批准号:14-050F);兰州理工大学校内基金

摘　　要：针对目前车牌定位算法准确率较低、定位速度慢等问题,提出一种基于颜色特征和改进Canny算子的车牌图像定位方法.该方法先利用彩色图像的颜色特征对车牌图像进行初步定位,提取该颜色特征时不包含亮度信息,有效克服了光照变化的影响;再对Canny边缘检测算法进行改进,提出一种新的梯度幅值和梯度方向一阶偏导数计算方法及高、低双阈值自适应确定方法;最后利用多组车牌图像对所提算法的性能进行验证,验证结果证明了所给方法的正确性和有效性.For the problems of low accuracy,and slow locating speed in the existing license plate locating algorithms,a method of license plate locating based on color feature and improved Canny operator was proposed.Firstly,the color feature of the image is used for license plate preliminary locating;the color feature extracted is not included in the luminance information,overcoming the effects of illumination changes effectively.Secondly,based on the in-depth research of the Canny edge detection algorithm,a new calculation method of first-order partial derivative is proposed for the gradient magnitude and gradient direction,at the same time,an adaptive method for high and low dual-threshold determination is also given.Many license plate images were used to evaluate the performance of the algorithm proposed,and the correctness and effectiveness were confirmed,which provides a basis for the pre-processing of license plate locating in the real-time license plate recognition systems used in the field of intelligent transportation.

关 键 词：车牌定位 颜色特征 CANNY算子 边缘检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]