基于暗通道的单幅图像除雾

针对暗通道先验算法存在的对全局大气光值估计不准确的问题,提出了一种单幅图像除雾的优化技术,对全局大气光值的选取方式进行了改进.首先用完美反射法对含雾图像进行颜色校正,接下来利用四叉树算法粗略估计大气光值,对其进行数值校正;再利用引导滤波对透射率进行优化,进而得到平滑的透射率.通过大气散射模型复原,得到较为清晰的除雾图像...

Single image defogging based on dark channel

Aiming at the problem of inaccurate estimation of the global atmospheric light value in the dark channel prior algorithm,a single image defogging optimization technique is proposed,which improves the selection of the global atmospheric light value.First,use the perfect reflection method to correct the color of the foggy image,then use the quadtree algorithm to roughly estimate the atmospheric light value and correct it numerically;therefore use the guided filter to optimize the transmittance to obtain a smooth transmittance.Through atmospheric scattering model restoration,a clearer defogging image is obtained.Finally,the image details are enriched by guided filtering.The experimental results show that the algorithm has high definition and good visual effects.