本文選題：圖像增強(qiáng)　切入點(diǎn)：低照度圖像　出處：《南京郵電大學(xué)》2017年碩士論文

【摘要】：在現(xiàn)實(shí)生活中,由系統(tǒng)采集設(shè)備所獲取的圖像和視頻,在周圍環(huán)境光照不足的情況下容易出現(xiàn)對比度下降、細(xì)節(jié)丟失、色彩失真等問題,這將嚴(yán)重影響到圖像的后續(xù)處理與應(yīng)用,因此有效地對低照度圖像進(jìn)行增強(qiáng)具有重要的意義。本文分析了不足光照環(huán)境下圖像降質(zhì)的原因及特性,研究常用的圖像增強(qiáng)相關(guān)算法,并根據(jù)實(shí)際情況對現(xiàn)有的算法進(jìn)行改進(jìn)和完善。本文的具體研究內(nèi)容如下:首先,本文針對采集的單幅低照度圖像提出了一種基于對數(shù)圖像處理模型的圖像增強(qiáng)算法。為了提高反射分量的提取效果,本文提出利用LIP模型改進(jìn)低通高斯濾波器來估計(jì)光照分量,更好地保持圖像的細(xì)節(jié)特征;為了進(jìn)一步提高圖像的對比度,防止局部過度增強(qiáng)現(xiàn)象,本文結(jié)合韋伯-費(fèi)希納定律提出了基于局部背景亮度信息的光照分量增強(qiáng)算法。實(shí)驗(yàn)結(jié)果表明,與傳統(tǒng)的增強(qiáng)算法相比,本文所提出的基于人眼視覺特性的增強(qiáng)算法,有效地提高了圖像的整體亮度和對比度,保留了圖像的細(xì)節(jié)信息,有較好的主觀視覺感受。其次,本文針對采集的低照度視頻提出了一種基于圖像融合技術(shù)的增強(qiáng)方法。針對低照度視頻圖像融合增強(qiáng)過程中容易產(chǎn)生圖像混淆、色彩漂移和前景運(yùn)動(dòng)區(qū)域增強(qiáng)效果不佳等問題,提出了一種改進(jìn)的基于高質(zhì)量視頻幀信息補(bǔ)償?shù)娜诤显鰪?qiáng)算法。該算法首先對低照度視頻中前景運(yùn)動(dòng)目標(biāo)和靜態(tài)背景進(jìn)行分離,對背景區(qū)域采用改進(jìn)的基于權(quán)重比例的融合策略來提高其質(zhì)量,對前景物采用限制對比度的直方圖均衡化對目標(biāo)進(jìn)行增強(qiáng),最后將背景與運(yùn)動(dòng)目標(biāo)融合得到最后的增強(qiáng)視頻幀。實(shí)驗(yàn)結(jié)果表明,本文提出的改進(jìn)融合策略的增強(qiáng)算法與傳統(tǒng)算法相比更好地凸顯圖像細(xì)節(jié)信息,亮暗區(qū)域的對比度都得到較大的改善,且不會(huì)出現(xiàn)混淆和色彩漂移等現(xiàn)象。最后,本文構(gòu)建了一種基于人眼視覺特性的圖像增強(qiáng)質(zhì)量評價(jià)方法。目前圖像增強(qiáng)缺乏有效統(tǒng)一的質(zhì)量評價(jià)方法,本文從人眼的主觀視覺特性出發(fā),提出了一種有效的低照度圖像增強(qiáng)效果評價(jià)方法,并建立相應(yīng)的評價(jià)函數(shù)。實(shí)驗(yàn)表明,該方法能夠?qū)Σ蛔愎庹窄h(huán)境下降質(zhì)圖像的增強(qiáng)效果給出較為準(zhǔn)確的評價(jià)。
[Abstract]:In real life, the images and videos captured by the system acquisition equipment are prone to the problems of contrast decline, detail loss, color distortion and so on, when the surrounding environment is not fully illuminated.This will seriously affect the subsequent processing and application of the image, so it is of great significance to enhance the low illumination image effectively.In this paper, the causes and characteristics of image degradation under insufficient illumination are analyzed, and the commonly used image enhancement correlation algorithms are studied, and the existing algorithms are improved and improved according to the actual situation.The main contents of this paper are as follows: firstly, an image enhancement algorithm based on logarithmic image processing model is proposed.In order to improve the extraction effect of reflection component, this paper proposes to use LIP model to improve the low-pass Gao Si filter to estimate the illumination component, to better preserve the detailed features of the image, and to further improve the contrast of the image.In order to prevent the phenomenon of local over-enhancement, this paper presents an algorithm of illumination component enhancement based on local background luminance information in combination with Weber-Fischner 's law.The experimental results show that compared with the traditional enhancement algorithm, the proposed enhancement algorithm based on human visual characteristics can effectively improve the overall brightness and contrast of the image, and retain the details of the image.Have good subjective visual feeling.Secondly, this paper proposes an enhancement method based on image fusion for captured low illumination video.Aiming at the problems of image confusion, color drift and poor performance of foreground motion region enhancement in the process of low illumination video image fusion enhancement, an improved fusion enhancement algorithm based on high quality video frame information compensation is proposed.The algorithm firstly separates the foreground moving target from the static background in the low illumination video, and improves the quality of the background region by adopting an improved fusion strategy based on the weight ratio.The object is enhanced by histogram equalization with restricted contrast. Finally, the final enhanced video frame is obtained by fusion of background and moving object.The experimental results show that the improved fusion algorithm presented in this paper can better highlight the details of the image compared with the traditional algorithm, and the contrast of the bright and dark areas will be improved greatly, and there will be no confusion and color drift.Finally, an image enhancement quality evaluation method based on human visual characteristics is proposed.At present, there is a lack of effective and uniform quality evaluation method for image enhancement. Based on the subjective visual characteristics of human eyes, this paper presents an effective method for evaluating the effect of low illumination image enhancement, and establishes the corresponding evaluation function.The experimental results show that the method can accurately evaluate the enhancement effect of degraded images in insufficient light environment.
【學(xué)位授予單位】：南京郵電大學(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6條

1 王小元;張紅英;吳亞東;劉言;;基于物理模型的低照度圖像增強(qiáng)算法[J];計(jì)算機(jī)應(yīng)用;2015年08期

2 肖進(jìn)勝;單姍姍;段鵬飛;涂超平;易本順;;基于不同色彩空間融合的快速圖像增強(qiáng)算法[J];自動(dòng)化學(xué)報(bào);2014年04期

3 盧振坤;;基于空域圖像局部多向梯度模的圖像融合方法[J];激光與紅外;2010年01期

4 張宇;付冬梅;李曉剛;孔維功;;基于特征的多層次紅外與可見光圖像融合方法[J];計(jì)算機(jī)應(yīng)用研究;2009年04期

5 王正友;黃隆華;;基于對比度敏感度的圖像質(zhì)量評價(jià)方法[J];計(jì)算機(jī)應(yīng)用;2006年08期

6 周旋,周樹道,黃峰,周小滔;基于小波變換的圖像增強(qiáng)新算法[J];計(jì)算機(jī)應(yīng)用;2005年03期

相關(guān)博士學(xué)位論文 前1條

1 饒?jiān)撇?夜間視頻增強(qiáng)的關(guān)鍵技術(shù)研究[D];電子科技大學(xué);2012年

，

本文編號：1729928