基于Matlab的车牌识别算法
基于Matlab的车牌识别算法
摘要
车牌系统是计算机视觉和模式识别技术在智能交通领域的重要应用课题之一。车牌识别系统是以特定目标为对象的专用计算机系统，该系统主要包括三个内容:车牌定位、字符分割和字符识别。 其中车牌定位的目的就是从所拍摄的汽车图像中确定车牌的位置,从而便于后续的字符分割和字符识别工作。目前常用的方法有：基于模板匹配的方法、基于特征的方法和神经网络法等。
本设计采用基于模板匹配算法和基于人工神经网络算法对车牌进行定位识别，此算法只对蓝底白字车牌进行分割识别，对黑底白字车牌原则上整个算法可直接适用，。此算法分割出的图像像素值和模板图像达到了一致，由此便避免了切割出的图像像素值不一致所带来的问题。但对白底黑字车牌、黄底黑字车牌，需要对车牌定位算法进行调整，并将图像反转（0变1、1变0）。
关键词：车牌识别系统； 字符分割； 车牌定位
1
基于Matlab的车牌识别算法
LICENSE PLATE RECOGNITION ALGORITHM BASED
ON MATLAB
ＡＢＳＴＲＡＣＴ
License plate system is a computer vision and pattern recognition technology in one of the important application research topic in the field of intelligent transportation. License plate recognition system based on specific goals of a special computer system, the system mainly includes three contents: license plate locating, character segmentation and character recognition. One of the purpose of license plate location is taken from the auto locate the license plate in the image, so as to facilitate the subsequent work character segmentation and character recognition. Now commonly used methods are: based on template matching method, based on the characteristics of the method and neural network, etc.
This design USES based on template matching algorithm and based on artificial neural network algorithm to locate license plate recognition, the algorithm is only for blue white license plate segmentation recognition, the algorithm can be directly applicable in principle to the black white plate,. This algorithm to segment the image pixel values and template image, thus to avoid the cut out in the process of image pixel values are not consistent. But black on white background and black text plate, yellow bottom plate, adjustments need to license plate localization algorithm, and the image inversion of (0, 1, 1, 0).
Key words: license plate recognition system; Character segmentation;
License plate location
2
基于Matlab的车牌识别算法
目录
1 前言 ........................................................................................................................................... 4
1.1 车牌号识别研究背景 .................................................................................................. 4 1.2 车牌号识别技术研究现状和趋势 ............................................................................ 5
1.2.1国内外车牌识别技术情况及我国车牌特点 ................... 5 1.2.2车牌识别技术的应用前景 ................................. 6 1.3 车牌识别研究内容 ............................................ 7 2 车牌识别系统设计原理概述......................................... 9 3 车牌识别系统程序设计............................................ 11
3.１ 图像读取及车牌区域提取 .................................... 11
3.１.1图像灰度图转化 ....................................... 11 3.１.2图像的边缘检测 ....................................... 13 3.１.3灰度图腐蚀 ........................................... 14 3.１.4图像平滑处理 ......................................... 15 3.１.5移除小对象 ........................................... 16 3.１.6车牌区域的边界值计算 ................................. 17 3.２ 字符切割 .................................................. 18
3.２.1字符切割前的图像去噪处理 ............................. 18 3.２.2字符切割前的图像膨胀和腐蚀处理 ....................... 19 3.２.3字符切割 ............................................. 19 3.３ 字符识别 .................................................. 22
3.３.1字符识别方法选择 ..................................... 22 3.３.2字符归一化 ........................................... 22 3.３.3字符匹配识别 ......................................... 23
4 仿真结果及分析.................................................. 26
4.1 车牌定位及图像读取及其图像处理 ............................. 26 4.2 车牌字符分割及其图像处理 ................................... 26 5 结论............................................................ 28 参考文献........................................................... 29 致谢............................................... 错误！未定义书签。
3