基于OpenCV与SVM的车牌识别方法  

License Plate Recognition Method Based on OpenCV and SVM

作　　者：聂文都 蔡锦凡 NIE Wendu;CAI Jinfan(Xijing University,Xi'an 710123)

机构地区：[1]西京学院,西安710123

出　　处：《计算机与数字工程》2021年第6期1244-1247,1268,共5页Computer & Digital Engineering

基　　金：陕西省教育厅科研项目(编号:16JK2237);西安学院研究生创新基金项目资助。

摘　　要：为了提高车牌识别效率,论文提出了一种基于OpenCV与SVM的车牌识别方法。用Canny算子获取图像边缘轮廓信息,基于SURF算法对车牌进行特征提取,并结合HSV颜色空间来定位车牌位置;基于垂直投影方法来提取车牌字符特征,再利用SVM算法实现对车牌字符的识别。该方法利用计算机视觉库OpenCV实现,利用OpenCV封装的SURF算法和SVM算法实现车牌识别,实验表明,该方法有较高的识别效率。In order to improve the efficiency of license plate recognition,a license plate recognition method based on OpenCV and SVM is proposed.Canny operator is used to obtain image edge contour information,SURF algorithm is used to extract features from the license plate,and HSV color space is combined to locate the location of the license plate.Based on the vertical projection method,the characters of license plate are extracted,and SVM algorithm is used to realize the recognition of license plate charac⁃ters.This method is realized by computer vision library OpenCV,and license plate recognition is realized by SURF algorithm and SVM algorithm encapsulated by OpenCV.Experiments show that this method has high recognition efficiency.

关 键 词：CANNY算子 SURF算法 HSV SVM OPENCV 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]