The Design of Grid-based Surveillance System with Embedded Ganglia Broker 陳家榮、高富建
E-mail: [email protected]
ABSTRACT
A traditional residence surveillance system only can continuously capture monitoring images from image sensor that cannot actively perform the intelligent identification. Image sensors produce lots of continuous monitoring images so that the system spends a lot of time and storage space in processing these images, which is needlessly expensive. The system captures all images from image sensors in turn, so some monitoring blind spots exist among image sensors in the traditional surveillance system. This research proposed a design of Grid-based Surveillance System with Embedded Ganglia Broker to improve the shortcoming of traditional surveillance systems in real-time monitoring. The proposed image sensor could judge whether images are abnormal or normal by comparing the difference of the background image and captured images. Consequently, the Embedded Ganglia Broker, which is designed by embedded system development board, would send those abnormal images to the adaptive back-end server with Grid Middleware ; therefore, this could prevent from storing a large number of images that are normal. The proposed intelligent image sensor also provides a mechanism for multiple front-end image sensors to capture the monitoring images in real time and solve the problem of monitoring blind spots in the traditional surveillance system.
In order to eliminate the congestion from images storing, this research proposed a distributed storing structure based on Grid technologies. The proposed system structure utilizes Embedded Ganglia Broker to integrate the backend-distributed computer resources and provide a real-time storage mechanism to solve congested problem caused by large images transmission. The proposed system also uses the RFT element of Globus tools for data replication to collect abnormal image files located at the back-end storing servers, and provides users to monitor and manage.
Key Words : Grid , Intelligent Image Sensor , Embedded System Keywords : Grid、Intelligent Image Sensor、Embedded System
Table of Contents
封面內頁 簽名頁 授權書 iii 中文摘要 iv ABSTRACT v 誌謝 vi 目錄 vii 圖目錄 ix
第一章 緒論 1 1.1 前言 1
1.2 研究動機與目的 2 第二章 嵌入式系統 4 2.1 嵌入式系統概述 4 2.1.1 嵌入式系統發展歷史 5 2.1.2 嵌入式系統的發展趨勢 6
2.2 本研究所採用之嵌入式系統開發板 7 2.3 DMA PXA270作業系統和編譯器 11 2.4 發展板開發所需套件工具 12 第三章 格網（GRID） 14 3.1 格網概述 14
3.2 格網架構 15
3.3 Grid Security Infrastructure（GSI） [5] 19
3.4 資源管? 20 3.5 資訊服務 21 3.6 資?管? 23
第四章 系統實作與分析 26
4.1 嵌入式影像儲存代理伺服器(EGB) 28 4.2 EGB實作 31
4.2.1 Ganglia Grid Portal 34 4.2.2 EGB 39
4.2.3 Qt 監控資料顯示 42
4.3 具格網功能之影像儲存伺服器設計與實作 43 4.3 EGB性能分析 44
第五章 結論 47
?考文獻 49
附錄A 安裝與建立Qt 桌面執行環境 51 附錄B Globus 安裝流程 57
附錄C平台移殖cross complier之建構 70 REFERENCES
[1]胡繼陽、李維仁、柯力群、張志龍，”嵌入式系統導論,3e”，學貫行銷股份有限公司，2004年。
[2]新華電腦股份有限公司，”ARM 內核嵌入式SOC原理以ARM7TDMI為例”，全華科投圖書股份有限公司，2004年。
[3]黃菖裕，『整合格網結構之智慧型監視系統設計』，私立大葉大學碩士論文，2008年。
[4]長高科技股份有限公司， http://www.dmatek.com.tw。
[5]劉家瑋，『智慧型居家保全系統之格網設計』，私立大葉大學碩士論文，2007年。
[6] http://www.uclinux.org/[7]Ian Foster, “What is the Grid? A Three Point Checklist”, Argonne National Laboratory & University of Chicage, 2002.
[8]Ian Forster, Carl Kesselman and Steven Tuecke, “The Anatomy of the Grid Enabling scalable Virtual Organizations”, Supercomputer Application, Page:2-6, 2001.
[9]The Globus Project, http://www.globus.org/[10] http://www.uclinux.org/[11] http://www.busybox.net/[12]
http://www.uclibc.org/toolchains.html[13]Embedded uClinux 在PreSOCes 上實作，新華電腦股份有限公司，全華科技圖書股份有限公司
，2004?。
[14]Ian Foster, Carl Kesselman and Steven Tuecke， “The Anatomy of the Grid Enabling Scalable Virtual Organizations” Supercomputer Application， Page:6-14， 2001.
[15] BORJA SOTOMAYOR, “ Globus Toolkit 4 PROGRAMMING JAVA Services” , Page:7-10，2005。
[16]OpenSSL, http://www.openssl.org/[17]Apache HTTP SERVER PROJECT, http:// httpd.apache.org/[18]
Http://www.globus.org/toolkit/docs/4.0/security/simpleca/
