本文選題：圖像配準　切入點：圖像融合　出處：《蘭州交通大學(xué)》2013年碩士論文

【摘要】：基于普通二維圖像的虛擬全景沉浸式漫游系統(tǒng)，與傳統(tǒng)的基于三維幾何建模的漫游系統(tǒng)相比，具有計算量少、獨立于場景、能達到照片級別清晰度、占用更少的存儲空間，不需要對計算機進行硬件提升，就能獲得真實感很強的實時交互響應(yīng)的視覺沉浸感�？蓮V泛應(yīng)用于虛擬政務(wù)場館、教學(xué)、房地產(chǎn)和旅游等行業(yè)的全景展示與漫游，是目前國內(nèi)外虛擬現(xiàn)實技術(shù)研究領(lǐng)域的技術(shù)熱點。 本文著重研究了普通二維圖像的虛擬漫游系統(tǒng)中所涉及的圖像配準、圖像融合等關(guān)鍵算法，并結(jié)合全景漫游系統(tǒng)中實景虛擬的構(gòu)建流程，以蘭州市旅游漫游系統(tǒng)的研發(fā)為基礎(chǔ)開發(fā)了基于全景圖像的虛擬漫游制作平臺。本文具體工作如下： (1)深入分析了當前全景漫游中圖像配準的通用算法，通過大量的研究發(fā)現(xiàn)：在處理圖像主體為對稱物體時，算法本身極易產(chǎn)生圖像特征誤匹配。本文提出了基于模糊聚類分析的SIFT（Scale Invariant Feature Transform）特征檢測算法，將模糊數(shù)學(xué)的模糊集與模糊矩陣引入到特征點的篩選當中，在此基礎(chǔ)上實現(xiàn)了全景圖像的精確配準，實驗證明該方法在處理全景圖像配準中配準效率高、算法魯棒性強。 (2)在對全景漫游系統(tǒng)中圖像融合算法研究的基礎(chǔ)上，針對圖像在融合過程中運動物體極易出現(xiàn)的鬼影、重影問題，及融合后圖像白平衡產(chǎn)生明暗不一致問題，結(jié)合局部優(yōu)先思想進行圖像多分辨率的縫合線融合，“點”及“線”、“線”及“面”最終生成一個的縫合區(qū)域，用于圖像融合。實驗結(jié)果表明加入局部優(yōu)先思想后算法可以有效的去除圖像融合過程中的運動鬼影和降低圖像白平衡問題，有效提升了圖像融合質(zhì)量。 (3)在研究實景空間虛擬漫游系統(tǒng)的建立技術(shù)與交互技術(shù)的基礎(chǔ)上，論文對視點空間漫游中的預(yù)測階梯過渡內(nèi)存調(diào)度算法與構(gòu)建圖像反投影柱面算法進行了優(yōu)化，在此基礎(chǔ)上研發(fā)了蘭州市旅游景點的虛擬漫游系統(tǒng)。
[Abstract]:Compared with the traditional roaming system based on 3D geometric modeling, the virtual panoramic immersion roaming system based on general two-dimensional images has less computation, is independent of the scene, can achieve photo level clarity, and occupies less storage space.The visual immersion of real-time interactive response can be achieved without hardware upgrading of the computer.It can be widely used in virtual government venues, teaching, real estate, tourism and other industries panoramic display and roaming, which is a hot spot in the field of virtual reality technology research at home and abroad.This paper focuses on the image registration, image fusion and other key algorithms involved in the general two-dimensional image virtual roaming system, and combines the construction process of the virtual scene in the panoramic roaming system.Based on the research and development of tourism roaming system in Lanzhou, a virtual roaming platform based on panoramic images is developed.The specific work of this paper is as follows:1) the general algorithms of image registration in panoramic roaming are deeply analyzed. Through a lot of research, it is found that when the image body is a symmetrical object, the algorithm itself is prone to produce image feature mismatch.In this paper, a feature detection algorithm of SIFT(Scale Invariant Feature transform based on fuzzy clustering analysis is proposed. The fuzzy set and fuzzy matrix of fuzzy mathematics are introduced into the feature points selection, and the panoramic image registration is realized.Experiments show that the proposed method is efficient and robust in panoramic image registration.2) based on the research of image fusion algorithm in panoramic roaming system, aiming at the problems of ghost, double image and white balance of image, which are easy to appear in the process of image fusion,Combined with the idea of local priority, the multi-resolution stitching line fusion of image is carried out, and the "points" and "lines", "lines" and "faces" are finally generated a suture region for image fusion.The experimental results show that the algorithm can effectively remove the moving ghost and reduce the white balance in the process of image fusion, and improve the quality of image fusion effectively.3) based on the research of the technology of virtual roaming system and interactive technology, this paper optimizes the scheduling algorithm of prediction step transition memory and the algorithm of constructing image backprojection cylinder in view space roaming.On this basis, a virtual roaming system of Lanzhou tourist attractions is developed.
【學(xué)位授予單位】：蘭州交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2013
【分類號】：TP391.41

【參考文獻】

相關(guān)期刊論文 前1條

1 金良海;李德華;;基于噪聲檢測的圖像去噪算法[J];模式識別與人工智能;2008年03期

，

本文編號：1719449