基于不均匀光照校正和优化透射率的夜间去雾

Nighttime Dehazing Based on Uneven Light Correction and Optimized Transmission

针对夜间有雾天气场景下的光照不均匀、图像不清晰问题,提出了一种基于夜间不均匀光照校正和优化透射率的去雾算法。首先针对光照不均匀问题,使用基于伽马函数的图像自适应校正算法,消除不均匀光照对图像的影响;其次,由于夜间图像大部分颜色较暗,通过改进局部大气光获得大气光值;利用Sigmoid函数融合亮通道与暗通道得到粗透射率,经过联合双边滤波细化,再对光源处透射率进行优化,并使用伽马函数增强得到最终透射率;然后根据大气散射模型得到初步恢复图像;最后经过引导滤波细化,得到最终图像。实验结果表明所提算法在夜间场景中能有效去雾,且能较好的保持图像颜色对比度。

In the hazy environment at night,the images captured by the observer are characterized by low contrast and blur caused by uneven ambient illumination,a defogging algorithm based on uneven illumination correction and transmission optimization at night is proposed.Firstly,for the problem of uneven illumination,an image adaptive correction algorithm based on Gamma function was used to eliminate the influence of uneven illumination on the image;secondly,since most of the night images are dark places,the atmospheric light value was obtained by improving the local atmospheric light;the rough transmission was obtained by fusing the bright channel and the dark channel through the Sigmoid function,after joint bilateral filtering refinement,the transmission at the light source was then optimized and the final transmittance was obtained using gamma function enhancement;then,the initial restoration image was obtained according to the atmospheric scattering model;finally,the guided filtering was refined to obtain the final image.The experimental results show that the proposed algorithm can effectively dehaze in the night scene,and can better maintain the image color contrast.