本文選題：低照度　切入點(diǎn)：鐵路　出處：《北京交通大學(xué)》2017年碩士論文

【摘要】：隨著城市規(guī)模的不斷發(fā)展,鐵路基礎(chǔ)設(shè)施陸續(xù)建設(shè)完成,列車技術(shù)取得了重大進(jìn)步,同時(shí)也對(duì)列車運(yùn)行安全提出了更高的要求。物體非法入侵鐵路軌間區(qū)域,引發(fā)鐵路交通事故,對(duì)鐵路的運(yùn)輸效率以及人民生命安全造成了極大威脅。在鐵路運(yùn)營中,通常需要使用鐵路眺望、圖像識(shí)別等方式探測軌間異物。由于低照度環(huán)境下采集的圖像對(duì)比度低、視覺效果差,細(xì)節(jié)信息無法清晰呈現(xiàn)出來,這使得后續(xù)的圖像特征提取、分析與識(shí)別受到限制。因此,對(duì)夜晚鐵路圖像進(jìn)行增強(qiáng)處理具有較強(qiáng)的實(shí)用性。論文分析了基于非物理模型和物理模型的低照度圖像增強(qiáng)算法,闡述了低照度圖像增強(qiáng)算法的基本過程并分析了其原理,結(jié)合夜晚鐵路環(huán)境,指出了現(xiàn)有"基于暗通道的低照度圖像增強(qiáng)算法"與傳統(tǒng)增強(qiáng)算法相比具有更高的保真度,對(duì)細(xì)節(jié)的保留效果更好。將現(xiàn)有"基于暗通道的低照度圖像增強(qiáng)算法"應(yīng)用于鐵路圖像中,出現(xiàn)曝光率過高以及強(qiáng)光源區(qū)域失效等問題,結(jié)合鐵路圖像特性從原理以及實(shí)驗(yàn)的角度對(duì)問題進(jìn)行剖析。本文主要研究的問題如下:第一,針對(duì)現(xiàn)有"基于暗通道的低照度圖像增強(qiáng)算法"在鐵路圖像強(qiáng)光源區(qū)域失效原因進(jìn)行分析,通過設(shè)定透射率閾值將強(qiáng)光源區(qū)域進(jìn)行圖像分割,對(duì)該區(qū)域使用局部方差對(duì)其進(jìn)行增強(qiáng)處理。第二,驗(yàn)證暗通道先驗(yàn)理論在軌間區(qū)域的適用性,并提出了適用于鐵路圖像的暗通道理論。根據(jù)鐵軌圖像的亮度和斜率特征,結(jié)合邊緣提取算法和Hough變化算法對(duì)軌間區(qū)域進(jìn)行截取,基于鐵路圖像的暗通道理論對(duì)夜晚軌間區(qū)域進(jìn)行增強(qiáng)處理。實(shí)驗(yàn)結(jié)果表明,本文算法可以有效提升夜晚鐵路圖像對(duì)比度,突出細(xì)節(jié)信息,從而獲得更加清晰、自然的高質(zhì)量圖像。
[Abstract]:With the continuous development of the city scale, the construction of railway infrastructure has been completed one after another, and the train technology has made great progress. At the same time, it has put forward higher requirements for the safety of train operation. Causes of railway traffic accidents, which pose a great threat to the efficiency of railway transportation and the safety of people's lives. In the operation of railways, it is usually necessary to use railways to look out. Image recognition and other methods are used to detect foreign bodies between tracks. Because of low contrast, poor visual effect and poor visual effect, the detailed information can not be clearly presented, which makes the subsequent image feature extraction, analysis and recognition limited. It is very practical to enhance railway images at night. This paper analyzes the low illuminance image enhancement algorithm based on non-physical model and physical model, expounds the basic process of low illuminance image enhancement algorithm and analyzes its principle. Combined with the night railway environment, it is pointed out that the existing "low illuminance image enhancement algorithm based on dark channel" has a higher fidelity than the traditional enhancement algorithm. The existing "low illuminance image enhancement algorithm based on dark channel" is applied to railway image, which has the problems of high exposure rate and failure of strong light source area, etc. Combined with the characteristics of railway images, the problems are analyzed from the angle of principle and experiment. The main problems of this paper are as follows: first, This paper analyzes the causes of the failure of the existing "low illuminance image enhancement algorithm based on dark channel" in the railway image strong light source region, and segments the strong light source region by setting the transmittance threshold. The local variance is used to enhance the region. Secondly, the applicability of the dark channel priori theory in the interorbital region is verified, and the dark channel theory suitable for railway image is proposed. According to the brightness and slope characteristics of the railway image, Combined with edge extraction algorithm and Hough variation algorithm, the interorbit region is intercepted and enhanced based on dark channel theory of railway image. The experimental results show that the proposed algorithm can effectively improve the contrast of the night railway image. Highlight the details to get a clearer, natural, high-quality image.
【學(xué)位授予單位】：北京交通大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：U298

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 陳爾言;王衛(wèi)靜;;基于導(dǎo)向?yàn)V波的暗通道去霧算法研究[J];科技展望;2016年04期

2 陳書貞;任占廣;練秋生;;基于改進(jìn)暗通道和導(dǎo)向?yàn)V波的單幅圖像去霧算法[J];自動(dòng)化學(xué)報(bào);2016年03期

3 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

4 黃勇;孫興波;袁文林;范云飛;;基于暗原色先驗(yàn)的低照度圖像增強(qiáng)[J];四川理工學(xué)院學(xué)報(bào)(自然科學(xué)版);2015年03期

5 韓正汀;路文;楊舒羽;劉奇;齊晶晶;;基于導(dǎo)向?yàn)V波優(yōu)化的自然圖像去霧新方法[J];計(jì)算機(jī)科學(xué)與探索;2015年10期

6 李睿;武曉春;;基于數(shù)字圖像處理的直線型鐵軌的自動(dòng)識(shí)別[J];電視技術(shù);2014年03期

7 劉洋;潘靜;龐彥偉;;一種基于暗原色先驗(yàn)的低照度視頻增強(qiáng)算法[J];電子測量技術(shù);2013年10期

8 王蓉;孫玉秋;;基于直方圖處理的局部圖像增強(qiáng)算法研究[J];長江大學(xué)學(xué)報(bào)(自然科學(xué)版);2013年01期

9 王智文;;幾種邊緣檢測算子的性能比較研究[J];制造業(yè)自動(dòng)化;2012年11期

10 譚艷;王宇俊;李飛龍;葛耿育;;幾種典型的圖像邊緣檢測算法的分析比較[J];電腦知識(shí)與技術(shù);2012年07期

相關(guān)博士學(xué)位論文 前1條

1 李艷梅;圖像增強(qiáng)的相關(guān)技術(shù)及應(yīng)用研究[D];電子科技大學(xué);2013年

相關(guān)碩士學(xué)位論文 前5條

1 卜令勇;低照度圖像增強(qiáng)算法研究[D];天津大學(xué);2014年

2 劉三亞;圖像局部增強(qiáng)算法的研究與實(shí)現(xiàn)[D];寧夏大學(xué);2014年

3 翟佳;基于暗原色先驗(yàn)的低照度視頻快速增強(qiáng)算法[D];天津大學(xué);2014年

4 姜雪松;不足光照條件下的圖像與視頻增強(qiáng)[D];哈爾濱工業(yè)大學(xué);2013年

5 丁銳;鐵路圖像分類和軌頂邊緣定位算法研究[D];西南交通大學(xué);2013年

，

本文編號(hào)：1673815