基于OpenCV的人脸检测与跟踪方法实现  

The Application of Face Detection and Tracking Method Based on OpenCV

作　　者：王蓉[1] 侯鹏鹏[1] 曾昭龙[1] 

机构地区：[1]中国人民公安大学警务信息工程学院,北京100038

出　　处：《科学技术与工程》2014年第24期115-118,126,共5页Science Technology and Engineering

基　　金：国家自然科学基金(71203229);北京高等学校青年英才计划项目(YETP1366);2014北京市支持中央高校共建项目公安视听专业建设;中央高校基本科研业务费专项基金(2014JKF01121)资助

摘　　要：人脸是视频内容中的重要信息,对人脸的检测与跟踪是智能视频分析的一个重要分支。实现了基于OpenCV的人脸检测与跟踪方法。首先计算视频图像的Haar-like特征,然后利用AdaBoost级联分类器进行人脸检测,最后使用基于颜色的Camshift算法实现人脸跟踪。在VS 2010开发平台上调用OpenCV函数库进行C++编程;并使用MFC框架实现了人脸检测与跟踪方法。仿真实验结果证明方法思路合理、计算复杂度较低、鲁棒性较好。Face is an important information in video,and dynamic face detection and tracking is an important part of video analysis and image recognition.This paper has achieved a method of the face detection and tracking based on OpenCV.The AdaBoost method is used to detect dynamic face,and the Camshift algorithm based on color is used to track dynamic face.The proposed method is programmed by using OpenCV library and C ＋＋ on VS 2010 development platform,and MFC framework is used to design the system interface.Simulating experimental results show that the face detection and tracking method in this paper is reasonable,which has a lower computational complexity and better robustness.

关 键 词：图像处理 人脸检测 人脸跟踪 OPENCV 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]