夜间有雾图像的光照模型构建及去雾

发布时间：2019-03-14 13:12

【摘要】：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。
[Abstract]:Fog images at night will be seriously degraded, and the presence of artificial light source also makes the ambient light appear non-uniform. In order to solve the above problems, this paper proposes an illumination model suitable for night fog images, and realizes the de-fog of night images on the basis of this model. Two parameters, ambient light and transmittance, are included in the model, both of which will change spatially with the local content of the image. Firstly, the parameters are initially estimated based on the information loss constraint theory, and then refined by fast guidance filtering to suppress block effect and halo effect. Finally, the refined parameters are replaced into the illumination model. The final target image to be restored can be obtained by solving the model. The experimental results show that the proposed algorithm can effectively remove fog at night, suppress the divergence of the bright area and reproduce the details of the dark area, and the recovered scene has better brightness and contrast. The restored image color is natural and the overall performance is better than other algorithms of the same type.
【作者单位】：西安理工大学自动化与信息工程学院;安康学院电子与信息工程学院;陕西省复杂系统控制与智能信息处理重点实验室;
【基金】：国家自然科学基金资助项目(No.61673318) 高层次人才科研启动项目(No.2016AYQDZR06)
【分类号】：TP391.41

【相似文献】