發(fā)布時(shí)間：2018-06-11 16:01

  本文選題：亮度圖像 + YIQ顏色模型�。� 參考：《西安理工大學(xué)》2016年碩士論文

【摘要】：戶外監(jiān)控系統(tǒng)對霧霾等極端天氣比較敏感,在霧霾天氣條件下,戶外視頻質(zhì)量出現(xiàn)嚴(yán)重退化,從而導(dǎo)致其戶外視頻監(jiān)控性能受到嚴(yán)重影響。因此,為了增強(qiáng)戶外視頻監(jiān)控的適用性和穩(wěn)定性,研究霧霾視頻圖像清晰化復(fù)原技術(shù)具有重要的實(shí)際應(yīng)用意義。本文的主要研究內(nèi)容是以單幅圖像去霧算法為基礎(chǔ),設(shè)計(jì)完整的適用于視頻監(jiān)控場景下的視頻去霧方法的方案,并使用霧天戶外監(jiān)控系統(tǒng)采集的視頻進(jìn)行驗(yàn)證,具體工作如下:1、針對暗先驗(yàn)去霧算法產(chǎn)生邊緣殘霧和運(yùn)行速度慢的問題,提出了一種基于亮度圖像的快速單幅圖像去霧方法。首先,對暗原色估計(jì)透射率的方法進(jìn)行改進(jìn),用YIQ顏色模型的亮度分量代替暗通道,并用中值濾波代替最小值濾波,以提高計(jì)算效率,并抑制光暈現(xiàn)象。其次,采用效率更高的四叉樹算法來求解大氣光值。最后,針對室外的各種有霧圖像進(jìn)行實(shí)驗(yàn)。實(shí)驗(yàn)結(jié)果表明,本文方法在處理速度及去霧效果上都優(yōu)于其它幾種經(jīng)典算法,且適用于視頻去霧。2、提出一種面向視頻監(jiān)控的圖像去霧方法。首先,根據(jù)監(jiān)控視頻背景基本不變的特點(diǎn),利用均值濾波的背景建模方法提取通用背景圖像;然后,利用上述所研究的單幅快速圖像去霧方法,基于所得到的通用背景圖像估計(jì)通用透射率圖和大氣光值;最后,利用大氣散射模型復(fù)原視頻的每一幀圖像,并將其連續(xù)輸出,得到清晰的視頻。實(shí)驗(yàn)結(jié)果表明,監(jiān)控視頻的去霧方法接近實(shí)時(shí)去霧的要求。3、在上述工作的基礎(chǔ)上,設(shè)計(jì)并實(shí)現(xiàn)了基于MATLAB7.9.0平臺的單幅圖像去霧系統(tǒng)和基于Visual Studio 2010和OpenCV 2.3.1開發(fā)的視頻去霧系統(tǒng)。該系統(tǒng)集成了本文所提出的去霧方法,以及部分經(jīng)典算法。
[Abstract]:The outdoor surveillance system is sensitive to extreme weather such as smog. Under the condition of smog, the outdoor video quality is degraded seriously, which results in its outdoor video monitoring performance being seriously affected. Therefore, in order to enhance the applicability and stability of outdoor video surveillance, it is of great practical significance to study the restoration technology of haze video image clarity. The main research content of this paper is to design a complete scheme for video de-fogging based on a single image de-fogging algorithm, and verify it by using the video collected by the fog outdoor monitoring system. The main work is as follows: 1. In order to solve the problem that dark priori defogging algorithm produces edge residual fog and slow speed, a fast single image de-fogging method based on luminance image is proposed. Firstly, the method of estimating the transmittance of dark primary color is improved. The luminance component of YIQ color model is used to replace the dark channel, and the median filter is used to replace the minimum filter in order to improve the calculation efficiency and suppress the halo phenomenon. Secondly, a more efficient quadtree algorithm is used to solve the atmospheric light value. Finally, experiments are carried out on various fog images outside. The experimental results show that the proposed method is superior to other classical algorithms in processing speed and de-fogging effect, and is suitable for video de-fogging. 2. A new image de-fogging method for video surveillance is proposed. First of all, according to the feature that the background of surveillance video is basically invariant, the background modeling method of mean filter is used to extract the general background image, and then the single fast image de-fogging method mentioned above is used. The general transmittance map and atmospheric light value are estimated based on the obtained general background image. Finally, each frame image of video is reconstructed by atmospheric scattering model, and the clear video is obtained by outputting it continuously. The experimental results show that the de-fogging method of surveillance video is close to the requirement of real-time de-fogging. Based on the above work, a single image de-fogging system based on MATLAB 7.9.0 and a video de-fogging system based on Visual Studio 2010 and OpenCV 2.3.1 are designed and implemented. The system integrates the de-fogging method proposed in this paper, as well as some classical algorithms.
【學(xué)位授予單位】：西安理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 周理;畢篤彥;何林遠(yuǎn);柏航;;基于人眼視覺機(jī)理的霧霾圖像質(zhì)量提升算法[J];中南大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年10期

2 何智文;陳巍;;基于Wiener濾波的快速視頻去霧算法[J];電視技術(shù);2015年15期

3 劉海波;楊杰;吳正平;張慶年;鄧勇;;基于暗通道先驗(yàn)和Retinex理論的快速單幅圖像去霧方法[J];自動化學(xué)報(bào);2015年07期

4 唐寧;呂洋;;基于物理模型的快速單幅圖像去霧方法[J];電視技術(shù);2015年09期

5 肖鐘捷;李寶方;;一種改進(jìn)的單幅圖像快速去霧方法與實(shí)驗(yàn)[J];地球信息科學(xué)學(xué)報(bào);2015年04期

6 於敏杰;張浩峰;;基于暗原色及入射光假設(shè)的單幅圖像去霧[J];中國圖象圖形學(xué)報(bào);2014年12期

7 吳迪;朱青松;;圖像去霧的最新研究進(jìn)展[J];自動化學(xué)報(bào);2015年02期

8 王一帆;尹傳歷;黃義明;王洪玉;;基于雙邊濾波的圖像去霧[J];中國圖象圖形學(xué)報(bào);2014年03期

9 張怡;張加萬;楊國強(qiáng);李亮;;基于時(shí)空一致性優(yōu)化的視頻去霧算法[J];計(jì)算機(jī)應(yīng)用研究;2011年10期

10 禹晶;徐東彬;廖慶敏;;圖像去霧技術(shù)研究進(jìn)展[J];中國圖象圖形學(xué)報(bào);2011年09期

，

本文編號：2005864