Nightti me Haze Re moval Based on Illu mination Esti mation

Shuai FANG,Yu-kun ZHAO,Xin-ke LI,Yong-jin LIU,Fei-ran JIE
DOI: https://doi.org/10.3969/j.issn.0372-2112.2016.11.002
2016-01-01
Abstract:Nighttime haze removal is more difficult,since nighttime hazy images contain degradation more serious than daytime’s due to illumination,resulting in lower brightness,lower contrast,uneven illumination,and chromatic aberra-tion,etc.Depending on imaging law of nighttime hazy image,a novel dehazing algorithm based on illumination estimation is presented.Firstly,the uneven illumination is removed by estimate the illumination map.Secondly,an entropy-based scheme for nighttime images is proposed to obtain transmission map.Finally,the color of illumination is estimated by statistics of source regions.The experimental results show that the proposed algorithm can effectively remove the uneven illumination effect,increase the image contrast and improve the image visual effect.
What problem does this paper attempt to address?