請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/66130
標題: | 除霧及去雨之數位影像視訊技術 Image Dehazing and Rain Removal Using Digital Image and Video Processing |
作者: | Tzu-Yen Lee 李慈晏 |
指導教授: | 貝蘇章(Soo-Chang Pei) |
關鍵字: | 除霧,除雨,影像復原,視頻復原,影像增強, Dehazing,Rain removal,Image Restoration,video Restoration,Image Enhancement, |
出版年 : | 2012 |
學位: | 碩士 |
摘要: | 在惡劣氣候下所造成的低能見度,是許多電腦視覺應用的主要問題,如監視,智能車輛,室外物體識別等。這是由於大氣微粒大量存在於參與介質(participating medium)中並且有顯著的分佈和大小時,所造成的現象。基於這一現象,可以將氣候狀況分為靜態和動態兩類。靜態的惡劣氣候是由很多微小的粒子所造成的,其通常在時間和空間上是一致的,如霧和霾;相反地,動態的惡劣氣候是由較大的粒子所造成的,如雨滴和雪花,其在分析上較靜態惡劣氣候困難,主要是因為被雨和雪影響的鄰近區域,在時間和空間上是不同的。這些狀況不僅會困擾和混淆人類的觀察者,還會降低那些依賴微小特徵的電腦視覺演算法之有效性,因此模擬在各種天氣下所造成的視覺影響和發展出一套良好的演算法來移除與消除這些由懸浮微粒所造成的影響,變得不可或缺與重要的!
在本篇論文中,我們首先介紹目前最具代表性的三種單一影像除霧的方法:contrast-based [1], independent component analysis [2], 和dark channel prior-based [3],並且基於dark channel prior-based的方法,我們更進一步的提出了一個有效且健全的方法來改善除霧的品質。相較於目前存在的方法,我們的方法除了可以對白天霧影像給予滿意的除霧品質,甚至對於夜晚霧影像,仍然可以提供良好的除霧結果。除此之外,我們還介紹目前最具代表性的三種去雨的方法:streak-based detection [4], image-based blurring [5], and frequency-based analysis [6]。接著,我們另外提出了一個簡單但有效的去雨方法,其主要的設計概念為結合average-based偵測雨滴和block-based移除雨滴的方法,是一個可以用較少的運算時間得到較高品質的去雨結果之方法。 Poor visibility in bad weather is a major problem for many applications of computer vision such as surveillance, intelligent vehicles, and outdoor object recognition, etc…. The reason is that the substantial presence of atmospheric particles has significant size and distribution in the participating medium. Based on this, weather conditions can be characterized as static and dynamic cases. Specifically, static bad weather such as fog and haze caused by microscopic particles are usually spatially and temporally consistent. Oppositely, dynamic bad weather has large particles such as raindrops and snowflakes. Because spatially and temporally neighboring areas are affected by rain and snow differently, the analysis is more difficult. Under these conditions, the human viewer would be annoyed and confused. They also degrade the effectiveness of any computer vision algorithm based on small features. Therefore, it is necessary to model the visual effects for the various cases and then remove them. In this thesis, we introduce three existing typical single image dehazing methods: contrast-based [1], independent component analysis [2], and dark channel prior-based [3]. To improve the dehazing quality, we propose a robust and effective dehazing method. Unlike other existing methods, our method gives satisfactory dehazing quality during daytime and nighttime. Besides, three existing typical rain removal methods: streak-based detection [4], image-based blurring [5], and frequency-based analysis [6] are also introduced in the literature. In this follows, we design a simple but effective rain removal method by combining the average-based rain detection and block-based rain removing procedures. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/66130 |
全文授權: | 有償授權 |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-101-1.pdf 目前未授權公開取用 | 5.46 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。