中图分类号：TP29 单位代码：10425
学号：G1205016
mntp
基于OpenCV的ARM嵌入式车牌识别系统的研究The Study of License Plate Recognition SystemBased on OpenCV And ARM
立磨衬板
学科专业：控制工程
研究方向： 检测技术与自动化装置
作者姓名： 蓝芸根
指导教师： 于佐军 副教授
二〇一五年十月
The Study of License Plate Recognition System Based on OpenCV And ARM
感应式垃圾桶A Thesis Submitted for the Degree of Engineering Master
Candidate: Lan YunGen
Supervisor：Associate Prof. Yu ZuoJun
College of Information & Control Engineering
portal认证服务器
China University of Petroleum (East China)
摘要
汽车的车牌号码如同是车辆唯一的身份证，而对于这种特殊身份的识别，使得它在车牌识别系统中起着至关重要的作用。 本文主要是基于最新的计算机视觉OpenCV开源图像处理库的基础上，针对于ARM 嵌入式平台进行移植，其中的工作内容包括车牌号码图像的采集及预处理、车牌区域定位、号码字符分割、号码字符识别、嵌入式平台的移植五个部分的关键技术，开发过程简洁、编程代码可移植到多平台系统实现车牌号码的识别。 论文的基本组成是首先对车牌图像进行处理，使用OpenCV相应的图像处理库编写代码，如灰度值化图像，适当调整车牌图像的明暗对比度，相应做出预处理，以提高图片的可识别性。同时根据车牌的几何特征，利用车牌的几何尺寸特征参数进行定位出车牌位置，之后运用车牌字符的基础知识对字符进行分割，而对于每个字符的识别过程中，可以利用BP神经网络算法对得到的大量车牌字符进行训练，以得到合适的权值和阈值，来对进行识别，最后将得到的车牌号码信息保存到文档且实现将编写好的识别应用程序移植到ARM平台中进行验证。
关键词：车牌识别，OpenCV，图像处理，嵌入式移植
The Study of License Plate Recognition System
pam加药
Based on OpenCV And ARM
LanY unGen（Control Engineering）
Directed by Associate Prof. YuZuoJun
Abstract
The car's license plate number is the only vehicle identification card, and for this special identification, making it plays a vital role in the license plate recognition system.
This paper is mainly based on the latest computer vision OpenCV image processing library，and based in ARM embedded platform transplant, which including image acquisition and pre processing, license plate location, character segmentation, character recognition, embedded transplantation five parts key techniques, and use the library OpenCV programming code to achieve the license plate automatic recognition.
After using the OpenCV corresponding image processing library to write code, through gray histogram statistics, and judge the value of the license plate image of the light and dark level, corresponding to make the pretreatment, in order to improve the image of the recognition. According
摄像头识别to the geometric features of the license plate, using Sobel operator to carry out the vertical direction of the texture, operation on the image, and then located the license plate. After the use of template matching and projection method, a large number of characters are trained by BP neural network. Finally, the identification of the application program is transplanted to the ARM platform for verification.
Key Word：V ehicle License Plate Recognition，OpenCV，Picture Processing，ARM
豫公网安备41160202000603
站长QQ:729038198
关于我们
投诉建议