摘  要

现代社会中企业、公司、工厂、机关单位等对于出入人员的身份验证越来越重要，对身份识别的准确性、高效性、便捷性、安全性、实用性提出了更高的要求。相比其他的生物特征识别技术而言，人脸识别具有其独特的优势,是最直接、最方便友好的方式，在人脸识别系统中主要涉及几个重要环节：人脸检测定位，人脸特征提取，分类识别。

本文重点分析当下学术界已有的基于模板匹配，规则特征搜索，人工神经网络等人脸检测方法以及基于向量机，特征脸，人工神经网络等人脸识别方法，结合实际需求，权衡准确性，高效性，实用性等，对比各种识别方法的优劣势，选取基于主成分分析（PCA）的特征脸方法作为本系统的核心算法。本系统在硬件构成上主要有三大块：OV511芯片的摄像头，博创up-techpxa270-s开发板与密码电子锁，工作流程如下:首先由摄像头抓取训练人脸图像库，当有人请求进入时，再有摄像头抓取人脸，采用基于特征脸的方法对该人脸进行身份识别，再根据识别结果开关电子锁。

该作品主要应用在公司，企业，机关单位等对出入人员需身份识别的领域，经验证本作品相对于目前已有的身份识别系统具有自然友好，方便快捷，高效准确等特点，具有良好的应用前景。

关键词：人脸识别；特征脸；门禁系统；QT界面

ABSTRACT

In modern society, enterprises, companies, factories, organs and units for the staff of identity verification is increasingly important, the identification accuracy, efficiency, convenience, safety, practicality put forward higher requirements. Compared to other biological feature recognition, face recognition has its unique advantages, It is the most direct, the most convenient and friendly manner, in face recognition system mainly involves several important aspects: face detection, feature extraction, classification and identification.

This paper focuses on the analysis of the current academic circles several existing face recognition methods, which are based on template matching, rule feature search, artificial neural networks and the method based on support vector machine, Eigenfaces method, combined with the practical demand, the trade-off between accuracy, efficiency, practicality, and compared the advantages and disadvantages of various identification methods, this system selection based on principal component analysis ( PCA ) method of eigenface as the system core algorithm. The system hardware composition has mainly three parts: OV511chip camera, up-techpxa270-s development board and the electronic cipher lock, the working process as follows: firstly, the camera grab training face image database, when a request for access, and camera crawl face, using the feature-based face method on the human face identification, then according to the result of recognition switch electronic locks.

This work mainly used in companies, enterprises, organs and units of staff required identification field, and this work compared to the existing identity recognition system has a advantage of natural, friendly, convenient, efficient and accurate, and has good application prospect.

Key Words: face recognition; eigenface; access control system; QT interface