基于android平台人脸检测与识别分析-电子与通信工程专业论文.docxVIP

摘要 摘要 摘要 随着手持移动设备的计算处理性能提高和存储空间增大，使得人脸检测识别 技术在手持终端上实现成为可能，继而完成自动个人身份认证功能。 本文主要研究基于 Android 系统平台的人脸自动检测和识别实现方法。首先 在 X86 平台上利用 VS2008 搭载 OpenCV 图形图像处理库对于 Haar 人脸检测方法 和 PCA 人脸识别方法进行了实验仿真，然后结合 Android 系统平台特性，对以上 两种算法进行了移植和优化，将带有人脸检测和识别接口的程序编译成*.so 库文 件供应用程序层调用，最终实现了通过 Android 系统手机摄像头获取图像，经过 系统 JNI 层完成 Haar 人脸检测和 PCA 方法人脸识别功能。 本文设计了 Android 平台自动人脸识别软件交互界面，在 Android 系统手机 中完成序列图像采集、人脸检测和人脸识别等功能。使用基于 Android 2.3 系统小 米一代手机（手机配置：内存 1G、双核 1.5GHz 主频）测试多幅 800*480 像素图 像，完成人脸检测（Haar 方法 150*150 检测窗口）和人脸识别（PCA 方法）平均 耗时 130ms，实现了对视频流中人脸的实时检测与识别，验证了本设计方法的可 行性和有效性，为进一步实现基于人脸识别技术的个人身份认证等相关应用打下 基础。 关键词：人脸检测 人脸识别 Haar 特征 主成分分析法 Android 系统 Abs Abstract Abstract With the improvement of the calculated performance and the increasing of the storage on the mobile device, face detection and recognition can be realized for identity authentication on these devices. The main research of this paper is how to release face detection and recognition automatically on the Android system. Firstly, this paper does the simulation experiment to accomplish face detection that is realized with Harr feature and face recognition that is realized with PCA algorithm on the PC with the OpenCV library. After learning the architecture of the Android system, the paper optimizes the program, and compiles the code of face detection and recognition interface to *.so library files which can be called by the application layer. Finally, the system gets images from the camera on the Android phone and the application realize face detection and recognition by calling the library through JNI method. This paper designs interactive interface of the automatic face recognition on the Android system. The function of the application includes collecting images, face detection and face recognition. The test is done on the first generation of the Xiaomi mobile phone (1G RAM, double cores, 1.5GHz) and the images resolution is 800*480. It realizes the face detection using Harr feature with 150*150 windows size and face recognition using P