ARM平台下人脸识别智能监控系统  

Implementation of intelligent face recognition surveillance system based on ARM

作　　者：江烂达 储珺[1] 缪君 JIANG Lan-da;CHU Jun;MIAO Jun(School of Software,Nanchang Hangkong University,Nanchang 330063,China)

机构地区：[1]南昌航空大学软件学院,江西南昌330063

出　　处：《计算机工程与设计》2018年第2期590-595,共6页Computer Engineering and Design

基　　金：国家自然科学基金项目(61663031;61661036);江西省研究生创新专项基金项目(YC2015-S337)

摘　　要：提出一种基于ARM的人脸识别智能监控系统设计方案。在嵌入式平台下,针对局部二值模式(local binary pattern,LBP)算法不能在较远距离快速实现人脸识别的问题,设计先利用ViBe(visual background extractor)算法实现背景建模,再对检测到的目标区域进行LBP人脸识别的方法。测试结果表明,相对于传统视频监控系统存在的效率低下、资源浪费、被动监控等问题,该系统功能完整、经济实用,能远距离快速识别非法闯入者,及时通知用户采取防范措施,具有广泛的应用前景。An intelligent face recognition video monitoring system based on ARM was proposed.Because of the limited operation capability of embedded platform,it is difficult for LBP algorithm to recognize face in real time from a long distance.To solve this problem,the system used the ViBe algorithm to detect moving object first,and then face was detected and recognized in the detected moving object.Compared with the traditional video monitoring system with low efficiency and poor intelligence,the system can recognize the illegal intruder in real time from a long distance,and remind the user to take preventive measures automatically.It is fully functional,economic,and with good practical application value.

关 键 词：嵌入式 人脸识别 智能监控 局部二值模式 运动目标检测 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]