随着计算机网络技术的快速发展及信息化进程的日益加快，信息安全和公共安全受到了越来越广泛的关注。由于准确的身份识别及鉴定是保证系统安全的重要前提，因此人脸识别技术越来越受到人们的青睐。它是一种基于人体生物特征的身份识别技术，是一个多学科交叉的边缘应用技术，它跨越了图像处理、模式识别、计算机视觉、生物学等学科。 本文首先对国内外关于人脸检测和人脸识别技术的常用算法进行了研究和分析。在对这些算法进行全面分析对比的基础上：在人脸检测方面，利用改进的Adaboost算法对通过摄像头采集的人脸进行检测，然后对采集到的人脸图像进行各种预处理，如人脸图像的光照增强、灰度变换、锐化、中值滤波、对比度增强、图像二值化进行预处理从而建立人脸库；在人脸识别方面，分别利用主成分分析法（PCA）和线性分析法（LDA）以及两者的改进和融合算法对ORL人脸库和自建人脸库进行分析及仿真实验；其仿真实验结果表明，上述改进算法不仅具有良好的检测与识别性能，而且具有较低的算法复杂度与便于硬件实现的优点。 在硬件实现方面，本文将上述算法代码移植到TMS-DM6446开发板中，并进行了优化。实验结果表明：该系统具有基本的人脸图像采集、预处理、检测与识别功能。

With the ever-accelerating process of the rapid development of computer network technology and information technology, information security and public safety has been more and more concerned. Accurate recognition and identification is an important prerequisite to ensure system security, therefore face recognition technology is welcomed by more and more people. It is an identification technology based on biometrics, is a multi-disciplinary Edge application technology which crosses the image processing, pattern recognition, computer vision, biology and other disciplines. This article carries out research and analysis of the commonly used algorithm of face detection technology at home and abroad. Based on the comprehensive analysis and comparison of these algorithms: At the aspect of face detection, do camera collected face detection using improved Adaboost algorithm and pre-treat the face images by the ways, such as high light of face images, gray-scale transformation, sharpening, median filtering, contrast enhancement, Image binarization pretreatment in order to establish the face database; in face recognition aspect, analysis and test the ORL face database and self-built face database using principal component analysis method (PCA), line analysis method(LDA) and the improvement and integration algorithm method, The result of simulation experiments shows that the improved algorithm not only has a good detection and recognition performance, and the algorithm has low complexity and advantages of ease hardware implementation. In hardware implementation aspect, the algorithm code is ported to the TMS-DM6446 devolopment board, and optimized.The experimental results show that: The system has a basic face image acquisition, preprocessing, detection and identification capabilities.