基于OMAP3530数字图像处理的多人脸识别系统设计  

Multi-face recognition system based on OMAP3530

作　　者：李硕豪[1] 张军[1] 

机构地区：[1]国防科学技术大学信息系统与管理学院,湖南长沙410073

出　　处：《微型机与应用》2013年第11期61-63,69,共4页Microcomputer & Its Applications

基　　金：国防科学技术大学创新型实践项目(2011020)

摘　　要：为了完成在不采用传感器条件下对公共场所人数的智能检测,从实际应用的角度出发,运用人脸图像拟合及模糊条件下人脸匹配的算法,实现了在嵌入式Linux平台下以OMAP3530芯片为核心的基于图像处理的多人脸识别系统。该系统能够得出公共场所实时人员的名单,并列出人员信息。以普通办公室为例进行实验,得出准确率可达80%。该系统充分利用了OMAP3530芯片强大的数据处理能力,具有性价比高、可维护性强、工作稳定的特点。In order to calculate the number of people intelligently in certain public place without sensor and realize multi-face recognition on the collection image, considering the practical application, this paper realizes the multi-face recognition system, by using ellipse fitting of face image and dim face recognition algorithm based on the image processing technology, under the embedded Linux platform with OMAP3530 chip as the core. This system can list the basic information of all those present. Experiments take genera/ office as an example and show that the rate of fight matching can achieve 80%. Taking full advantage of the OMAP3530 chip＇s powerful data processing ability, this system is possessed of high quality-price rate, easy maintenance and stabilization.

关 键 词：LINUX OMAP3530 数字图像处理 多人脸识别 

分 类 号：TN948.61[电子电信—信号与信息处理]