论文总字数：22471字

摘 要

如今高校宿舍的人员进出主要依赖人工登记,遇到外来用户进入宿舍,不能立刻确认他的身份,极大地降低了宿舍的安全性。近几年大学生宿舍的盗窃、伤人等事件频有发生,如果有一套可以确认来访人身份的系统，那么这个问题会得到有效地解决。有些高校采取了使用学生卡作为进入宿舍楼的方式,但该方法存在明显弊端，一是学生卡作为唯一凭证这种方式并没有很高的安全性,二是需要比较重的硬件设施(主要是闸门),增加了学校的支出。所以，本课题提出开发一套轻便的高校宿舍门禁系统。

本系统分为web软件系统和硬件系统两个部分。Web系统部署在远程服务器上，硬件系统由树莓派4B为主控制板,提供了led灯,音响作为输出模块，用来提醒进入的学生能否进入，摄像头作为输入模块，用来采集人脸信息以及人脸识别。Web软件系统，是一具有前端和后端的网站，设有两层权限，第一层是宿舍管理员，负责给他所管理的宿舍楼进行学生信息录入，查看学生基本信息，查看学生的进入记录等，系统中有多个宿舍管理员。第二层是系统管理员，负责管理所有的宿舍管理员。硬件树莓派中,主要包括三个部分，人脸采集,人脸识别,硬件模块响应。第一，人脸采集部分,在用户管理的添加学生基本信息时,由树莓派启动人脸录入程序，将学生的人脸录入到系统中，并传给Web网站；第二，人脸识别部分，基于OpenCV，采用LBPH算法，在光线良好的条件下，识别率可达到90%；第三，硬件模块响应部分，将人脸识别的结果通过调用硬件输出模块响应给用户。

关键词： 高校宿舍门禁管理；人脸识别；OpenCV；Raspberry Pi

College Students Dormitory Access Control System Based on Raspberry Pi

Abstract

Nowadays, college students' dormitories are mainly registered manually by the dormitories. When foreign users enter the dormitories, they cannot immediately confirm their identity, which greatly reduces the security of the dormitories. In recent years, thefts, injuries and other incidents have occurred frequently in college students' dormitories. If a system is set up to confirm the identity of visitors, this problem can be effectively solved. In the process of investigation, I found that other schools used student cards as the way to enter dormitory buildings. On the one hand, the student cards as the only voucher did not provide high security; on the other hand, heavy hardware facilities (mainly gate) increased the expenditure of the school. Therefore, the development of a set of portable dormitory access control system, has become my goal.

This system is divided into Web system and hardware system. The Web system is placed on the rented server. The hardware system is dominated by Raspberry Pi 4, which provides LED lights, audio as the output module to remind students whether they can enter, and camera as the input module to collect face information and face recognition. In the Web system, set up a website with front end and back end, the website is divided into two layers of permissions, the first layer is the dormitory administrator, responsible for his management of the dormitory building students input, view the student information, view the student's entry record. There are multiple hostel administrators in the system. The second level is the system administrator, who is responsible for managing all the dormitory administrators. In Raspberry Pi, it mainly includes three parts: face acquisition, face recognition and hardware module response. Face acquisition part, user management, now add the student's name, student id, dormitory, into the face, from the drop-down box to school students student id, click on the photo, raspberry pie face entry procedures, to enter the student's face into the system, and passed on to the site, the dormitory administrator user can check to add or delete student information change. Face recognition part, the use of OpenCV's own LBPH algorithm, the set of system in good light conditions, the recognition rate can reach 90%, in the ordinary light environment also reached 70%, the calculation results through the logic call hardware module response to the user. In this paper, the construction and realization process of the system are described in detail, and the expected goal is achieved.

Key words:Raspberry Pi; Face Recognition; OpenCV; Python

目录

摘 要 I

Abstract： II

第一章 绪论 1

1.1 引言 1

1.2 课题研究背景与意义 1

1.3 人脸识别发展概况 1

1.4 人脸识别前景与弊端 2

1.5 本文研究内容 4

1.6 论文组织结构 4

第二章 相关技术介绍 5

2.1 树莓派 5

2.2 OpenCV 5

2.3 人脸检测算法 6

2.4 三种面部识别算法 6

2.5 mqtt通信协议 8

2.6 本章小结 9

第三章 宿舍门禁硬件设计 10

3.1 主控模块 10

3.2 检测模块 10

3.3 电源模块 11

3.4 报警模块 12

3.5 体温传感模块 12

3.6 本章小结 13

第四章 宿舍门禁软件设计 14

4.1 安装树莓派系统 14

4.2 配置树莓派中文环境 15

4.3 树莓派GPIO介绍 16

4.4 树莓派连接 17

4.5 本章小结 18

第五章 宿舍门禁主要功能实现 19

5.1 web系统 19

5.2 SQLite3 22

5.3 MQTT 22

5.4人脸检测与录入 23

5.5人脸识别 24

5.6 本章小结 25

第六章 系统测试 26

6.1 硬件测试 26

6.2 软件测试 27

6.3 本章小结 28

第七章 结束语 29

致 谢 30

参考文献 31

第一章 绪论

1.1 引言

随着经济水平的不断发展，人们的生活的提高，同时随之而来的安全问题成为了关注的重点。由原先的数字密码验证，刷卡验证，现在又出现了许多指纹验证，语音验证，人脸验证。可以知道未来生物识别会得到广泛的应用^[1-2]。

现如今我国微电子产业不断有新的突破，图像识别技术也不管出现新的技术，使得面部识别的发展有了更低的实现成本，这就使得这项技术有了大众化的资本。面部识别技术最近几年发展十分迅速，不仅用于手机解锁，手机支付，而且在门禁方面也有很大的用处。这项技术的应用使得门禁系统，不再使用传统的手段，输入密码或者使用门禁卡，而是使用便携性更强，安全性更高的面部识别，使得面部识别这项技术在以后的生物识别技术中会占有一席之地。所以本次毕业设计选取面部识别的应用领域之一—门禁领域^[2-5]。

1.2 课题研究背景与意义

利用数字密码或者某些电子卡来充当个人身份的方法快要过时了，因为人们发现这种方法很容易被破解，而且不容易保管。新的时代下产生了新的识别方式——生物识别。这项技术一经推出，便受到了人们的热捧，更是被一度认为成未来的识别。原因是显而易见的，那就是生物识别的不可替代性和便携性。

剩余内容已隐藏，请支付后下载全文，论文总字数：22471字