影像強化與去模糊之研究

機構典藏 > College of Engineering > Graduate Institute & Department of Computer Science and Information Engineering > Thesis > Item 987654321/105720

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/105720

Title:	影像強化與去模糊之研究
Other Titles:	Study on image enhancing and deblurring
Authors:	莊驊;Chuang, Hua
Contributors:	淡江大學資訊工程學系碩士班林慧珍;Lin, Hwei-Jen
Keywords:	影像去模糊;影像強化;邊緣強化;環效應;粒子群最佳化演算法;反卷積;image deblurring;ringing artifacts;Image enhancement;edge enhancement;Particle Swarm Optimization (PSO);deconvolution
Date:	2015
Issue Date:	2016-01-22 15:03:33 (UTC+8)
Abstract:	隨著時代的演進，數位設備的取得已相當普遍，影像的應用也有廣泛被應用的趨勢，影像的品質很可能因為環境或設備的不良因素而受到影響，因而有許多改善影像品質的方法與技術被提出，如影像強化、影像去模糊、影像去雜訊、影像超解析等。其中影像強化主要是解決影像取像時因強光、光線不足或拍攝方式不當等不良因素造成影像過亮、過暗、或雜訊產生等問題。A. Gorai and A. Ghosh [1]提出了一種基於啟發式的演算法針對影像亮度和對比度做調整與強化，其對一般光線上有問題之下取得的影像都可得到不錯的強化結果；但是針對失焦、手振、物體移動、環境霧氣等所造成的影像模糊並沒有很好的去模糊效果。針對此類模糊影像大部分的方法都是利用模糊核來進行反卷積來達到去模糊的目的，其能有效地得到去模糊結果。本論文之研究分二個部分：（一）影像強化：改進A. Gorai and A. Ghosh提出的目標函數（objective function）與轉換函數（transformation function）進而改進其影像強化之結果，（二）影像去模糊：對模糊影像提出模糊核估測方法，再以估測出的模糊核對模糊影像進行去模糊。 A. Gorai and A. Ghosh [1]提出的影像強化方法根據一個目標函數（objective function）當適應度（fitness）利用粒子群優化（Particle Swarm Optimization, PSO）對一個影像轉換函數（transformation function）求出最佳參數組，再以求得的轉換函數對影像做轉換，得到一個影像強化結果。本論文將改進他們所提的objective function 與 transformation function進而改進其影像強化之結果。影像去模糊一般分為“盲去模糊法”與“非盲去模糊法”二種。在去模糊的過程中若無已知的模糊核，必須自行估測模糊核，此稱為盲去模糊；反之若具備已知模糊核則稱之非盲去模糊，本篇論文所提的影像去模糊方法屬盲去模糊。清晰影像中的邊緣具有較分明的顏色分佈，根據這個特性我們提出了一個邊緣強化的濾波器使模糊影像達到這個效果得到參考影像，再利用此參考影像來預估模糊核，最後再以取得的模糊核對模糊影像做反卷積運算得到估測清晰影像。 Multimedia is ubiquitous and the application of digital imaging is prolific, yet environmental conditions and hard ware limitations may adversely affect image quality. Advanced techniques such as image enhancement, deblurring, denoise, and super resolution have been developed to improve image quality post-digitization. Image enhancement is primarily concerned with problems caused by overexposure, underexposure, poor photographic technique, and optical noise. A. Gorai and A. Ghosh [1] proposed a method based on a heuristic algorithm to enhance images by adjusting brightness and contrast. Lighting problems are resolved effectively through this method, yet there are limitations when applied to blurred images, i.e., images with defocus blur, motion blur, handshake blur, or fog blur. For those kinds of blurred images, we need a specific technique of image deblurring to apply. As a result, this research is primarily concerned with (a) image enhancement: improving upon the objective function and transformation function proposed by A. Gorai and A. Ghosh and (b) image deblurring: a proposed method to estimate blur kernel and its application towards image deblurring. The proposed image enhancement algorithm improves the contrast and properly adjusts brightness of an image separately. Each part is achieved by a PSO algorithm. Based on whether the blur kernel is known or not, the problem can be categorized into non-blind deblurring and blind deblurring. Unblind deblurring is applied to an image with a known blur kernel whilst blind deblurring is applied to an image without any given blur kernel. This paper proposes a blind deblurring method needing to predict a blur kernel in our own way. The color distribution of edge is more distinct in a clear image than in a blurred image. A filter is proposed to make edges in a blurred image clearer for use as a reference image. The blur kernel is estimated from this reference image. The blurred image is then deconvolved with the estimated blur kernel to introduce a latent image.
Appears in Collections:	[Graduate Institute & Department of Computer Science and Information Engineering] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	148	View/Open

Loading...