Fast single image dehazing based on improved dark channel prior

To solve the issues of unclear edges, low image contrast, and slow computation speed when dehazing using dark channel prior, an improved fast image dehazing algorithm based on dark channel prior was proposed. First, a hierarchical search algorithm based on quadtree decomposition was employed to estimate the atmospheric light, which can avoid objects with brighter colors than the atmospheric light being selected as reference for the atmospheric light. Second, the atmospheric scattering model was employed to obtain the initial transmittance, and then introducing fast guided filtering with edge information to refine it, which improves the issues of unclear edges and low contrast. Finally, the gamma correction algorithm was used to adjust the brightness and contrast of the dehazing image. By comparing experimental results on outdoor hazy images with other dehazing algorithms, it is demonstrated that the proposed method can effectively maintain clear edges, enhance image contrast, and accelerate dehazing speed.