基于ARM嵌入式人脸识别系统的设计和实现
摘要本课题旨在基于ARM设计并实现一个嵌入式人脸识别系统。课题工作首先基于C++语言实现了人脸识别算法中的人脸检测、图像预处理及特征提取等主要功能模块。其次在ARM开发板上通过移植Linux操作系统、QT图形化界面库和OpenCV计算机视觉库，搭建起识别算法所需的嵌入式平台。然后针对嵌入式系统对人脸识别算法进行了改写和移植。最后编写相应的嵌入式程序，支持系统通过摄像头捕捉人脸图像并完成识别功能。实验结果表明，所设计的嵌入式在光照适宜及无遮挡物的情况下能够达到令人满意的识别效果。嵌入式的人脸识别系统能够弥补一般的识别系统不易携带的缺点，能够预见拥有广阔的应用前景。73274
毕业论文关键词 嵌入式 人脸识别 交叉编译 目标检测
毕业设计说明书外文摘要
Title The design and implementation of an ARM-based embedded face recognition system
AbstractThe research task of this paper is to design and implement a face recognition algorithm on an ARM-based embedded system。 This project starts with implementing the basic functional blocks of face recognition algorithm using C++ programming, including face alignment, face image preprocessing and characteristic extraction。 The next step is to establish the embedded platform for recognition algorithm, through transplanting Linux operating system, QT graphical interface library and Open Source Computer Vision Library on the ARM development board。 Subsequently, this project transplants the face recognition algorithm for the purpose of compatibility with the established embedded platform。 The last step is to design an embedded program that supports the designed system in capturing face images by camera and accomplishing the task of face recognition。 Experimental results show that the face recognition algorithm achieves a satisfactory recognition accuracy under the circumstances of suitable lighting and no obstructions。 The embedded face recognition system by this work overcomes the shortcoming of lack of portability that exists in normal face recognition systems, and is expected to have a brilliant prospect in various applications。
Keywords Embedded systeam Face recognition Cross compiation Object detection
目 次
1 引言 1
1。1 研究背景与意义 1
1。2 相关研究现状 1
1。3 本课题主要工作 2
1。4 本文结构安排 2
2 人脸识别技术综述 3
2。1 人脸识别概论 3
2。2 人脸识别方法简介 3
2。3 人脸识别技术预备知识 3
2。4 本章小结 5
3 人脸识别主要算法实现 6
3。1 人脸识别算法概论 6
3。2 人脸识别算法具体实现 8
3。3 本章小结 13
4 嵌入式系统平台搭建与实现 14
4。1 嵌入式操作系统选取 14
4。2 开发环境介绍 14
4。3 环境配置与平台搭建流程