由于雾霾天气下空气中介质粒子的影响，成像设备所捕获的图像通常会存在对比度低、色彩丢失等问题。针对这些问题，本文提出一种基于雾气分布的大气光幕估计去雾算法。首先，将Fade方法提取的初始雾气分布图像进行阈值分割与细化处理得到较为准确的雾气分布信息。其次，通过分析景深、雾浓度、大气光幕之间的关系建立大气光幕估计模型。最后，通过亮度权重图与自适应大气光阈值选取大气光区域，获取较为准确的大气光，进而恢复退化场景。实验与结果表明，本文算法恢复的图像清晰自然，去雾效果彻底，并且能够保留图像中的细节信息。

Due to the influence of medium particles in the air under hazy weather, the images captured by imaging devices usually suffer from low contrast and color loss. To address these problems, an atmospheric light veil estimation dehazing algorithm based on haze distribution is proposed in this paper. Firstly, the initial haze distribution image extracted by twelve features of the image is subjected to thresholding and refinement to obtain more accurate haze distribution information. Secondly, the atmospheric light veil estimation model is established by analyzing relationships among the depth of field, haze concentration, and atmospheric light veil. Finally, the atmospheric light region is selected by luminance weighting map and adaptive atmospheric light threshold, and then recover the degraded scene. The experimental results show that this algorithm restores images clearly and can retain detailed information in the image.