【摘要】：隨著我國網(wǎng)絡(luò)時代的到臨,越來越多的場合會用到圖像處理技術(shù),其中,在車牌識別,人臉識別,視頻監(jiān)控等領(lǐng)域作用越來越明顯。這就要求在具體實(shí)踐的過程當(dāng)中充分還原真實(shí)的圖像以利于識別,但是在實(shí)際情況當(dāng)中,當(dāng)遇到霧霾天氣,由于空氣中的大氣粒子的散射作用,將導(dǎo)致所拍攝出來的圖像變得模糊,對比度與色彩等視覺效果也將大打折扣。因此,在處于惡劣天氣下,如何去除霧霾所帶來的影響,提高圖象質(zhì)量便顯得尤為重要,這也是本文所討論的重點(diǎn)。本文討論了兩類圖像去霧技術(shù),一類以圖像增強(qiáng)為主,圖像的增強(qiáng)技術(shù)主要表現(xiàn)在圖像的空域增強(qiáng)、頻域增強(qiáng)以及基于色彩恒常性的Retinex增強(qiáng)算法,其中重點(diǎn)討論并在不同情境下實(shí)驗(yàn)了基于Retinex的多種圖像去霧算法;另外一類以圖像復(fù)原技術(shù)為主,通過對圖像降質(zhì)模型的分析,以及對大量無霧圖像的統(tǒng)計(jì)與實(shí)驗(yàn),形成了基于暗原色先驗(yàn)的圖像去霧技術(shù)。本文細(xì)致分析了霧霾的形成原因以及具體的圖像降質(zhì)模型,并通過實(shí)驗(yàn)驗(yàn)證基于暗原色先驗(yàn)的圖像去霧技術(shù)的效果,對實(shí)驗(yàn)結(jié)果中表現(xiàn)出的時效性不足的問題進(jìn)行了修正,采取了一種基于引導(dǎo)濾波的透射率優(yōu)化算法提高了算法效率。最后本文通過對圖像降質(zhì)模型的的分析與研究,采取了基于類高斯濾波及自適應(yīng)中值濾波的快速去霧算法,不僅提高了去霧之后圖片的質(zhì)量,與暗原色先驗(yàn)算法相比提高了圖像成像的速度。本文各算法的模擬實(shí)驗(yàn)均基于MATLAB 2012下運(yùn)行,對于最終的實(shí)驗(yàn)結(jié)果通過對比度、平均梯度、信息熵以及EPI這四個量化參數(shù)進(jìn)行具體對比,并分析各算法的優(yōu)劣。
[Abstract]:With the advent of the Internet era in China, image processing technology will be used in more and more occasions, in which, the role of license plate recognition, face recognition, video surveillance and other fields is becoming more and more obvious. This requires that the real images be fully restored in the process of concrete practice to facilitate identification, but in practice, when the haze weather is encountered, due to the scattering of atmospheric particles in the air, The image will be blurred and visual effects, such as contrast and color, will be greatly compromised. Therefore, in severe weather, how to remove the impact of haze and improve the image quality is particularly important, which is also the focus of this paper. In this paper, two kinds of image de-fogging techniques are discussed, one is image enhancement, the other is image enhancement in spatial domain, frequency domain and Retinex enhancement algorithm based on color constancy. Retinex-based image de-fogging algorithms are discussed and tested in different situations. The other is based on the image restoration technology. Through the analysis of the image degradation model and the statistics and experiments of a large number of fog-free images, an image de-fog technique based on a priori dark primary color is formed. In this paper, the causes of haze formation and specific image degradation model are analyzed in detail, and the effectiveness of image de-fogging technology based on dark primary color priori is verified by experiments, and the lack of timeliness in the experimental results is corrected. A transmission optimization algorithm based on guided filter is adopted to improve the efficiency of the algorithm. Finally, through the analysis and research of image degradation model, a fast de-fogging algorithm based on Gao Si-like filter and adaptive median filter is adopted in this paper, which not only improves the quality of image after de-fogging, but also improves the image quality. Compared with the dark primary color priori algorithm, the image imaging speed is improved. The simulation experiments of each algorithm in this paper are based on MATLAB 2012. The final experimental results are compared by four quantization parameters: contrast, average gradient, information entropy and EPI, and the advantages and disadvantages of each algorithm are analyzed.
【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 楊梅;彭九慧;;霧霾天氣成因分析及應(yīng)對思考[J];農(nóng)業(yè)科技與信息;2016年26期

2 宋曉敏;趙紅東;盧俏;夏士超;席瑞媛;李夢宇;肖夢琪;;霧霾天氣下降質(zhì)圖像的清晰化處理[J];電訊技術(shù);2016年02期

3 汪東芳;鄭睿;;淺析圖像快速去霧與清晰度回復(fù)技術(shù)[J];信息與電腦(理論版);2016年02期

4 徐琳;陳強(qiáng);汪青;;色彩熵在圖像質(zhì)量評價中的應(yīng)用[J];中國圖象圖形學(xué)報(bào);2015年12期

5 何艷;方帥;;一種局部多尺度retinex算法在霧天圖像中的應(yīng)用[J];合肥工業(yè)大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年10期

6 王園宇;劉鈸;劉杰;張海超;;基于粉塵濃度的圖像退化模型[J];機(jī)械工程與自動化;2015年05期

7 盧彥飛;張濤;鄭健;李銘;章程;;基于局部標(biāo)準(zhǔn)差與顯著圖的模糊圖像質(zhì)量評價方法[J];吉林大學(xué)學(xué)報(bào)(工學(xué)版);2016年04期

8 何寧;王金寶;鮑泓;;單幅圖像去霧方法研究綜述[J];北京聯(lián)合大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年03期

9 曾浩;尚媛園;丁輝;周修莊;付小雁;;基于暗原色先驗(yàn)的圖像快速去霧[J];中國圖象圖形學(xué)報(bào);2015年07期

10 趙長霞;段錦;李光明;彭杰;;基于大氣散射模型的偏振圖像去霧方法[J];長春理工大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年03期

相關(guān)碩士學(xué)位論文 前2條

1 王奕權(quán);圖像去霧與圖像增強(qiáng)算法研究[D];南京郵電大學(xué);2015年

2 劉安娜;霧天彩色圖像復(fù)原方法研究[D];中國海洋大學(xué);2014年

，

本文編號：2456490