This article proposes an optimized image dehazing algorithm that combines sky region segmentation , DCP(Dark Channel Prior) and BCP (Bright Channel Prior) to address the problems of image distortion, decreased contrast, high saturation, and poor processing of sky regions in traditional dehazing methods. Firstly, the Shannon entropy of the gradient map of a foggy image is obtained, which leads to the texture map of the foggy image. The texture map can roughly determine the location of the sky region. Combined with the region growth algorithm and Canny, the sky region is finely segmented, and the atmospheric light value is calculated within the final obtained sky region. In addition, this article introduces the BCP and estimates the transmission through the fusion of dark and bright channels. Finally, a clear image is obtained based on the Atmospheric scattering model. The experiment shows that the proposed dehazing algorithm can effectively optimize the problem of traditional dehazing algorithms failing in the sky region, and the obtained dehazing images are more satisfying for human visual experience. At the same time, there are also good objective evaluation indicators that verify the feasibility, effectiveness, and superiority of the proposed algorithm.