基于統(tǒng)計(jì)量的圖像去霧算法研究
[Abstract]:Image processing aims to highlight some details in the image for visual observation and computer analysis. Under the condition of haze, the visibility of outdoor images is restricted and the contrast of the images is decreased due to the action of atmospheric particles. In order to solve the problem of image quality degradation, most of the current processing schemes are based on image enhancement and image restoration de-fog algorithm, image enhancement algorithm by improving the contrast of the image to achieve clarity effect; Based on the foggy image imaging model, the atmospheric scattering mechanism is modeled to restore the possible clear image without fog. Based on the fog image formation model and degradation mechanism, this paper explores the key techniques and implementation methods of image de-fogging, and gives the prior statistical evidence of dark channel and the improved non-local de-fogging algorithm. The main work of this paper is as follows: (1) the dark channel priori is a statistical rule based on the clear outdoor fog-free images, that is, there are at least one pixel with low intensity of color channel in the non-sky region of most outdoor fog-free images. In this paper, we assume that the three channels are independent of each other, and the scene points are independent with other pixel points in the field. The RGB value of the color fog free image is considered as a statistical variable. Assuming that these three variables are all subject to the Beta distribution, the density function and distribution function of the variables are given after the minimum filtering of two times (the RGB three-channel is small, then the RGB is small in a neighborhood). In order to verify the validity of dark channel priori. (2) the existing image de-fogging methods can be divided into local and non-local categories according to the different prior information used. Berman et al., based on the non-local clustering characteristics of clear images in RGB space, The geometric representation of a fog line (Haze-Line) for each color class of a fog image is constructed. The maximum radiative coordinates of the fog line (LRC:Largest Radial Coordinate) are the key to estimate the initial transmittance. In this paper, an unbiased estimate of LRC is given from the point of view of statistics. The experimental results show that the proposed method can obtain at least the same results as the original method.
【學(xué)位授予單位】:寧夏大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2017
【分類號】:TP391.41
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 孫曉曉;楊峰;;基于小波變換融合技術(shù)的去霧方法[J];山東師范大學(xué)學(xué)報(bào)(自然科學(xué)版);2016年02期
2 王質(zhì)春;俞文燕;;單幅霧霾天氣交通監(jiān)控圖像去霧處理技術(shù)[J];交通運(yùn)輸研究;2016年02期
3 畢篤彥;葛淵;李權(quán)合;任志河;南棟;陳劍鵬;;單幅圖像去霧方法研究[J];空軍工程大學(xué)學(xué)報(bào)(自然科學(xué)版);2013年06期
4 李利榮;汪蒙;;一種高效的圖像增強(qiáng)去霧算法[J];湖北工業(yè)大學(xué)學(xué)報(bào);2013年05期
5 劉佳嘉;周超;;一種基于色彩恒常理論的薄霧圖像增強(qiáng)方法[J];科學(xué)技術(shù)與工程;2013年13期
6 馬云飛;何文章;;基于小波變換的霧天圖像增強(qiáng)方法[J];計(jì)算機(jī)應(yīng)用與軟件;2011年02期
7 黃黎紅;;一種基于單尺度Retinex的霧天降質(zhì)圖像增強(qiáng)新算法[J];應(yīng)用光學(xué);2010年05期
8 張麗;楊安洪;李新濤;王建榮;;海島區(qū)域影像薄霧處理模型與實(shí)現(xiàn)[J];測繪科學(xué)技術(shù)學(xué)報(bào);2010年04期
9 汪榮貴;楊萬挺;方帥;吳昊;;基于小波域信息融合的MSR改進(jìn)算法[J];中國圖象圖形學(xué)報(bào);2010年07期
10 葉秋果;宗景春;李釧;滕惠忠;;基于同態(tài)濾波的遙感影像去云霧處理[J];海洋測繪;2009年03期
相關(guān)博士學(xué)位論文 前3條
1 趙宏宇;霧天圖像清晰化技術(shù)的研究[D];北京工業(yè)大學(xué);2015年
2 胡學(xué)友;霧天降質(zhì)圖像的增強(qiáng)復(fù)原算法研究[D];安徽大學(xué);2011年
3 翟藝書;霧天降質(zhì)圖像的清晰化技術(shù)研究[D];大連海事大學(xué);2008年
相關(guān)碩士學(xué)位論文 前7條
1 劉超;Curvelet變換在圖像去霧處理中的應(yīng)用研究[D];云南大學(xué);2016年
2 賈冬雪;圖像去霧算法的研究與應(yīng)用[D];沈陽工業(yè)大學(xué);2015年
3 代永珍;單幅有霧圖像去霧算法研究[D];國防科學(xué)技術(shù)大學(xué);2014年
4 劉明;基于MSR改進(jìn)算法的圖像和視頻去霧研究[D];安徽大學(xué);2014年
5 王一涵;霧天圖像增強(qiáng)方法研究[D];西安電子科技大學(xué);2012年
6 劉建明;基于PC的霧天圖像和視頻的清晰化算法研究與實(shí)現(xiàn)[D];上海交通大學(xué);2009年
7 祝培;惡劣天氣環(huán)境下圖像的清晰化[D];西安理工大學(xué);2004年
,本文編號:2347391
本文鏈接:http://sikaile.net/kejilunwen/ruanjiangongchenglunwen/2347391.html