English  |  正體中文  |  简体中文  |  Items with full text/Total items : 27454/39300
Visitors : 2534828      Online Users : 29
RC Version 4.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
LoginUploadHelpAboutAdminister

Please use this identifier to cite or link to this item: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/35742

Title: 基於改良型粒子群演算法之數位影像穩定研究
Investigation of the Digital Image Stabilization Based on Modified Particle Swarm Optimization Algorithm
Authors: Hong-Kuan Yu
余泓寬
Contributors: NTOU:Department of Communications Navigation and Control Engineering
國立臺灣海洋大學:通訊與導航工程學系
Keywords: 數位影像穩定;移動向量估測;移動向量補償;眾數;多目標粒子群演算法
Digital image stabilization;Motion vector estimation;Motion vector compensation;Mode;Multi-objective particle swarm algorithm
Date: 2013
Issue Date: 2013-10-07T02:58:27Z
Abstract: 一般的攝影機在拍攝時,會因為人手的抖動、車子行駛顛簸的震動、風力的吹動以及其他外力干擾下而使拍攝出來的影像序列會產生不必要的抖動。本論文擬提出基於數位PI控制器之數位影像穩定(Digital Image Stabilization;DIS)技術用以去除不必要的抖動影響以保留原影像的真實性。所提影像穩定技術主要分為全域移動向量,即移動向量估測單元與防止抖動又可讓影像序列平滑移動的移動補償向量,即移動補償單元。 移動向量估測方面則先利用Sobel邊緣偵測法找出整張影像中物件的邊緣,並根據物件的邊緣最明顯的部分,進而訂定出特徵區塊;接著在特徵區塊內求得區域移動向量,再以眾數法從區域移動向量中求得全域移動向量,此全域移動向量就是實際的影像移動向量。而移動向量補償方面則以平滑指標(Smoothness Index;SI)和移動誤差(Moving Error;ME)來評估影像穩定系統性能的好壞;同時亦以多目標粒子群演算法(Multi-Objective Particle Swarm Optimization;MOPSO)的適應值用於數位PI控制器參數的調整依據,進而達到讓影像序列可以在不同干擾環境下既可防止抖動又可平滑移動的效果。經手持實測的實驗結果顯示所提之數位影像穩定系統,除能有效的防止影像序列的晃動外,並且能在影像平移時消除影像延遲的缺點。
General camera while shooting, because of manpower jitter, car driving Britain slope vibration, wind blowing, and other external interference leaving the shooting out of the image sequence will produce unwanted jitter. This thesis intends to propose a digital PI-based digital image stabilization technology to remove unwanted jitter effect such that the authenticity of the original image can be preserved. Proposed image stabilization technology includes the global mobile vector (the motion vector estimation unit) and the motion compensation vector (the motion compensation unit) which can prevent jitter and allow smooth movement of the image sequence. First the Sobel edge detection method will be applied to find the edges of objects in the entire image for the motion vector estimation, and the feature block will be established according to the most visible part of the object edges. Then the feature area within the block move vector can be obtained and the global motion vectors can be derived from the local motion vectors using the mode algorithm. The global motion vector should be the actual image motion vector. For the motion vector compensation, the smoothing index (SI) and moving error (ME) will be used to evaluate the performance of the image stabilization system. Meanwhile, the parameters of the digital PI controller will be adjusted according to the fitness value of the multi-objective particle swarm algorithm such that the effects of the jitter prevention and smooth movement for the image sequences can be achieved in different interference environments. The handheld experimental results show that the proposed digital image stabilization system not only can prevent the judder of the image sequence effectively but also eliminate the image delay due to the image translation.
URI: http://ethesys.lib.ntou.edu.tw/cdrfb3/record/#G0010067037
http://ntour.ntou.edu.tw/handle/987654321/35742
Appears in Collections:[通訊與導航工程學系] 博碩士論文

Files in This Item:

File Description SizeFormat
index.html0KbHTML115View/Open


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

 


著作權政策宣告: 本網站之內容為國立臺灣海洋大學所收錄之機構典藏,無償提供學術研究與公眾教育等公益性使用,請合理使用本網站之內容,以尊重著作權人之權益。
網站維護: 海大圖資處 圖書系統組
DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback