English  |  正體中文  |  简体中文  |  Items with full text/Total items : 6498/11670
Visitors : 27687848      Online Users : 88
RC Version 3.2 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
Login ‧ Upload ‧ Help ‧ About ‧ Administer

Please use this identifier to cite or link to this item: http://ir.ncue.edu.tw/ir/handle/987654321/11853

Title: 應用競爭性類神經網路融合數值資料與影像資訊於追蹤系統之研究
The Research of Applying Competitive Hopfield Neural Network to Fuse Quantity Data and Image Information to Tracking System
Authors: 鍾翼能;胡永柟
Contributors: 電機工程學系
Keywords: Quantity data and image information;Data association technique;Competitive Hopfield Neural Network
Date: 2010-08
Issue Date: 2012-07-02T02:13:20Z
Publisher: 行政院國家科學委員會
Abstract: 雷達追蹤系統中，多目標追蹤(Multiple-Target Tracking, MTT)是不可缺少的重要技術。追蹤系統是否能夠正確的預估目標物的真實軌跡其中牽涉兩個問題:資料結合(Data Association technique)與變速度(Maneuvering)檢測。本研究提出運用競爭式類神經網路(Competitive Hopfield Neural Network, CHNN)特殊的運算架構，發展一資料相關結合運算的程序，以輔助雷達追蹤系統。本研究同時運用影像處理技術，考慮目標物的形態與外觀等，提高目標物被辨認的機率，降低預估的誤差。影像經前處理後，接著擷取目標物的特徵，並運用相似度函數進行影像特徵辨識，最後，以結構相似度(Structural Similarity, SSIM)，求得目標物的影像數值資訊，同時配合CHNN 資料結合技術與適應性程序追蹤輔助架構，能更準確追蹤目標。經模擬結果顯示，本研究提出的演算法能在複雜的追蹤環境中，有效降低追蹤多個變速度目標時的追蹤誤差。
Multiple-target tracking (MTT) is a prerequisite step for radar surveillance systems. Data association is the key technique in a radar multiple -target tracking system. A new approach to data association using both quantity data and image information is investigated in this project. In order to combine two different attributes, a fusion algorithm based on the Competitive Hopfield Neural Network (CHNN) is developed to match between sensor measurements and existing target tracks. When target maneuvering problems are occurred, an adaptive multiple-model maneuvering estimator is applied. Based on the computation algorithm, we convince that this approach can successfully solve the multiple -target tracking problems and have better performance.
Relation: 計畫編號:NSC99-2221-E018-009; 研究期間:201008-201107
Appears in Collections:[電機工程學系] 國科會計畫

Files in This Item:

File SizeFormat
2050200712012.pdf144KbAdobe PDF460View/Open


All items in NCUEIR are protected by copyright, with all rights reserved.

 

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