夜间有雾图像的光照模型构建及去雾

夜间有雾图像会出现严重退化,而且人工光源的存在也使得环境光呈现不均匀性。针对上述问题,本文提出了一种适用于夜间有雾图像的光照模型,并在此基础上实现了夜间图像去雾。模型中主要包含了环境光和透射率两个参数,这两个参数都会随着图像局部内容的变化而产生空间变化。首先基于信息损耗约束理论对上述参数进行初始估计,随后利用快速导向滤波对其进行细化,以抑制块效应和光晕效应,最后将细化后的参数代入光照模型中,通过求解模型即可获得最终待还原目标图像。实验结果表明,本文提出的算法能够有效实现夜间有雾图像的去雾处理,在抑制亮区发散的同时能重现暗区的细节,恢复的场景具有较好的亮度和对比度,恢复的图像颜色自然,总体性能优于同类型的其它算法。

Lighting model construction and haze removal for nighttime image

Nighttime haze images degrade seriously and artificial light sources cause uneven atmospheric light.Aimed at those problems,this paper introduced a kind of lighting model applicable to nighttime haze image and realizes nighttime image dehazing.The model mainly consisted of two parameters of atmospheric light and transmission,of which space changed with change of local image content.Firstly,the above parameters were initially estimated based on information loss constraint theory and then refinement was carried out by applying fast guide filter to restrain the block effect and halo effect;finally,ultimate refined parameters were applied to lighting model and target image could be obtained by solving the model.Experimental result shows that the proposed algorithm can realize dehazing treatment for nighttime haze image and show again detail of dark space by suppressing emission of light area simultaneously;recovered scene has favorable brightness,contrast ratio and recovered image has natural color.The overall performance of the algorithm is superior to that of other algorithms.