淡江大學機構典藏:Item 987654321/103166
English  |  正體中文  |  简体中文  |  Items with full text/Total items : 55184/89457 (62%)
Visitors : 10680565      Online Users : 16
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/103166


    Title: 多攝影機監視系統
    Other Titles: Surveillance System Based on Multi-Camera Network
    Authors: 顏淑惠
    Contributors: 淡江大學資訊工程學系
    Keywords: 監測;多攝影機;連續性可適的平均位移法;監測視區;超解析;surveillance;multi-camera;CamShift Algorithm;field of view (FOV);super-resolution
    Date: 2012-08
    Issue Date: 2015-05-20 09:30:41 (UTC+8)
    Abstract: 多攝影機監視系統Surveillance System based on Multi-Camera Network 我們提出一個兩年期的計畫案。此計畫是建立一個無重疊區的多攝影機監視系統。第一年著重於物件追蹤與多攝影機間的關係。第二年則是著重於目標物超解析以及系統實際架設與調整。 我們利用類似中間值濾波器的方法很快速的建立背景影像以達到目標物的擷取。接著以CamShift演算法來作物件追蹤。不同的高斯模組事先根據目標物的行動方向、速度、色彩資訊加以訓練,以預測目標物於不同攝影機視區(FOV)之間的對應關係。 對於目標物的超解析,我們採取一整合傳統方法與基於範例的方法。此一方法不需要很多的訓練影像,它藉由我們所選出有代表性的幾張影像以及這些影像的降維影像裡學習高/低解析的關係,達到具有效率與品質的高解析影像。最後,我們將系統實際架設於淡江校園的工學院大樓以擔任安全監控的任務,同時檢測、調整與改進的我們系統。
    Surveillance System based on Multi-Camera Network 多攝影機監視系統 We propose a two year project to construct a surveillance system based on non-overlapping Multi-Camera Network. The first year focuses on multi-camera geometry and object tracking; the second year focuses on target super-resolution and system unifying and tuning. In object detecting, the system uses an approximated median filter to modulate and update the background. CamShift algorithm is then adopted for object tracking. Various pre-trained Gaussian models are used for object matching between different field of views (FOVs). Models include the motion direction of the object, the velocity of the object, the color information of the object. To accomplish the super-resolution, we apply a hybrid method integrating classical and example-based methods. Especially, this method does not require a large training database. It learns the low-resolution/high-resolution relation among multiple scales of a few representative images. Finally, the system will be set up in the Engineering building of our campus to enforce the mission of surveillance.
    Appears in Collections:[Graduate Institute & Department of Computer Science and Information Engineering] Research Paper

    Files in This Item:

    There are no files associated with this item.

    All items in 機構典藏 are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - Feedback