基于最小值通道与对数衰减的图像融合去雾算法

雾天各类图像采集系统获取的图像颜色退化，细节模糊，严重影响户外成像系统的稳定性和有效性，因此研究图像去雾技术很有必要。针对暗通道一类去雾算法边缘去雾不彻底问题，提出一种基于最小值通道与对数衰减的融合去雾算法。首先，对有雾图像的最小值通道图进行对数衰减作为先验假设条件，再进行交叉双边滤波消除纹理效应，在操作前后分别进行下采样和上采样操作以提高运算速度，求出初始透射率；然后，用Canny算子检测最小值通道图得到的边缘进行对数衰减，得到边缘信息图，将初始透射率与边缘信息图进行加权融合构成优化透射率；最后，结合改进的四叉树搜索法求得的大气光值反解大气散射模型，恢复无雾图像。实验结果表明:所提算法可以有效抑制光晕现象，去除边缘残雾，且实时性好。 

Image fusion dehazing algorithm based on minimum channel and logarithmic attenuation

The image color degradation and blurred details acquired by various image acquisition systems in foggy weather seriously affect the stability and effectiveness of outdoor imaging system, so it is necessary to study image dehazing technology. Aimed at the incomplete edge dehazing of dark channel dehazing algorithm, a fusion dehazing method based on the minimum channel and logarithmic attenuation is proposed. Firstly, the logarithmic attenuation of the minimum channel map of the foggy image is taken as a priori hypothesis, and then cross-bilateral filtering is performed to eliminate the texture effect. Before and after the operation, downsampling and upsampling operations are performed respectively to improve the operation speed, then we get the initial transmittance. Secondly, Canny operator is used to detect the edge of the minimum channel and logarithmic attenuation is carried out to obtain the edge information map. The initial transmittance and the edge information map are weighted and fused to compose the optimal transmittance. Finally, the atmospheric light value obtained by the improved quadtree search method is used to solve the atmospheric scattering model and restore the fog-free image. The experimental results demonstrate that the proposed algorithm can effectively suppress halo effect and remove edge residual fog, and has good real-time performance.