基于透射率估计模型和引导滤波的雾气图像去雾算法

针对雾气条件下成像设备采集图像退化严重的问题，提出一种雾气图像的去雾算法。通过对雾气天气成像物理模型的简化，找到图像复原函数中的透射率和大气光值2个关键未知量；分析影响透射率的因素，通过对大量雾气图像进行灰度分布概率统计，提出一种透射率快速估计算法；通过引导滤波估计大气光值，利用简化的修复函数完成对雾气图像的去雾处理。与HE算法(HE Kaiming, SUN Jian, TANG Xiaoou. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2011, 33:2341)相比较，经本算法去雾复原后的图像信息熵值最少可提高0.060 4比特/像素、平均梯度值最少可提高0.009 55。实验结果表明，经本算法复原的有雾图像清晰度较高，细节复原较好，去雾效果明显。

Algorithm About Fog Image Dehazing Based on Transmittance Estimation Model and Guided Filtering

In order to solve the problem of serious degradation of image acquisition equipment under the condition of fog, a fog image defogging algorithm is proposed in this paper presents. Fog weather imaging model was simplified and 2 key variables (light transmittance and atmospheric transmittance value) in the image restoration function was found. We analyzed the influence factors on refractive index as well as the statistics of image gray distribution and a fast estimation algorithm of the model combined with the transmittance was put forward. This algorithm is utilized to estimate atmospheric light by guiding filtering and perform the processing of fog fog image by simplifying repair function. Compared with the HE algorithm (HE Kaiming, SUN Jian, TANG Xiaoou. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2011, 33:2341) the image after recovery by this defogging algorithm can improve 0.060 4 bit / pixel by using the measure method information entropy, and improve 0.009 55 by using the measure of the standard minimum average gradient. The experimental results show that the proposed algorithm has a better effect on the restoration of the fog image, and the recovery of the details. The effect of defogging is significant.