嵌入式车牌图像拍摄与识别节点设计

论文总字数：31479字

目 录

1绪论 3

1.1研究背景和意义 3

1.1.1智能交通系统发展现状 3

1.1.2车牌识别技术概述 3

1.1.3嵌入式车牌识别的优点 4

1.2国内外车牌识别技术研究现状 4

1.2.1国内外相关研究概况 4

1.2.2中国车牌识别技术的特殊性 4

1.3设计研究的内容及论文结构 4

1.3.1设计研究内容 4

1.3.2论文结构 5

2基于图像处理的车牌识别原理 6

2.1车牌识别算法组成 6

2.2车牌定位算法及原理 6

2.2.1基于颜色特征的车牌定位算法 6

2.2.2基于纹理特征的车牌定位算法 7

2.2.3基于边缘特征的车牌定位算法 7

2.3 字符识别算法及原理 8

2.3.1字符分割算法原理 8

2.3.2字符识别算法分类及原理 9

2.4本章小结 9

3车牌定位原理及实现 10

3.1车牌定位概述 10

3.2图像预处理 10

3.2.1高斯模糊降噪处理 11

3.2.2图片灰度化处理 11

3.2.3 Soble算子边缘检测处理 12

3.2.4图片二值化处理 13

3.2.5形态学操作 14

3.3图像分割截取 15

3.3.1取轮廓截取ROI感兴趣区域处理 15

3.3.2对感兴趣区域ROI图块的判断处理 16

3.4本章小结 17

4字符识别原理及实现 17

4.1字符识别概述 17

4.2车牌图像预处理 17

4.2.1二值化、灰度化处理 17

4.2.2消除边框、铆钉的影响 18

4.3车牌图像字符分割及识别 18

5 SVM车牌判断及ANN字符判断原理和实现 19

5.1 SVM支持向量机和ANN人工神经网络介绍 19

5.1.1 SVM支持向量机简介 19

5.1.2 ANN人工神经网络简介 20

5.1.3 SVM与ANN的联系与区别 20

5.2 SVM支持向量机进行车牌判断 21

5.3 ANN人工神经网络进行字符判断 21

5.4本章小结 22

6嵌入式车牌识别系统的实现 22

6.1 PC机模拟车牌识别环境的搭建 23

6.1.1选择Linux Ubuntu系统及安装 23

6.1.2 Unbuntu下配置OpenCV 24

6.1.3 Ubuntu下安装Qt软件开发工具包 24

6.2 ARM开发板的选择 25

6.3 ARM开发板的使用 27

6.3.1重新烧写树莓派3b开发板系统 27

6.3.2实现树莓派开发板远程登陆与远程桌面控制 27

6.3.3开发板与PC机如何进行文件互传 28

6.4树莓派开发板开发环境的搭建及程序的移植 29

6.4.1开发环境搭建 29

6.4.2树莓派开发板上摄像头的安装与使用 29

6.4.3程序的移植 30

6.5本章小结 31

7总结与展望 32

7.1总结与展望 32

参考文献: 32

致谢 34

嵌入式车牌图像拍摄与识别节点设计

孙毅

，China

Abstract：Intelligent transportation system is the direction of the continuous development of modern transportation system. The so-called intelligent transportation system is the integration of the latest computer technology, information technology, electronic communication and other technologies into the modern transportation system. Through intelligent transportation systems, we can establish efficient, stable and safe transportation systems. By managing the license plate, the vehicle can be effectively managed, so the license plate recognition technology can be said to be one of the core of the intelligent transportation system. Aiming at the unfavorable features such as large volume of license plate recognition and installation trouble, an ARM-based embedded license plate recognition system was designed to solve the problem of portability. This system first simulates the x86 framework program from the PC Linux Ubuntu system, and then uses the powerful performance of various types of ARM development boards, such as the Raspberry Pi development board, to directly establish the same development environment as the PC in the ARM development board. The program is transplanted, and the program that generates the ARM framework runs on the ARM development board. It can recognize the license plate photograph taken by the camera and take out the character information of the license plate portion and the license plate in the original image. And because it uses Qt development, it has a friendly human-computer interface. The research content of this design for license plate recognition technology includes the two-step operation of license plate positioning and character recognition. The license plate positioning includes pre-processing, positioning, and interception of the originally captured picture, and the input SVM support vector machine determines whether the output is license plate after normalization processing. The character recognition includes preprocessing the license plate image output by the SVM, and inputting the ANN artificial neural network to determine the character after the character is captured. The entire system first uses Qt to develop simulations based on the open source OpenCV library in the Linux system in the VMware virtual machine, and then migrates to the ARM development board to run under the Ubuntu system to implement embedded development. Because this system is developed with Qt, it is easy to transplant. After being transplanted into the ARM development board, it has advantages such as convenience, speed, stability, and low power consumption.

Key words：license plate location; character recognition; Linux; OpenCV; embedded design

剩余内容已隐藏，请支付后下载全文，论文总字数：31479字