A Dark Channel Prior Dehazing Algorithm Combining Mean Curvature and Histogram Analysis
A dark channel prior dehazing algorithm that combines mean curvature and histogram analysis is proposed to address the issues of edge blur,oversaturation,and misestimation of atmospheric light values in current dark channel prior algorithms.Firstly,calculate the mean curvature of all pixels on the micro geometric surface of the input image,and fuse the small-scale and large-scale transmission maps with the normalized mean curvature intensity as the weight;secondly,analyze the histograms of the three channels of the foggy image to determine whether there are overly bright areas and estimate the global atmospheric light value;finally,a clear dehazing image is obtained using the atmospheric scattering model.The experimental comparison with other algorithms on public datasets such as HAZERD shows that this algorithm can solve the problems of current dark channel algorithms and has higher robustness and effectiveness.
国家科技期刊平台
NSTL
万方数据
维普
