Natural and optical remote sensing image dehazing based on BL-G-BM3D decomposition model

Atmospheric conditions such as suspended particles and haze can significantly affect the contrast and color of the natural and optical remote sensing images. In this paper, we introduced the variational image decomposition technique into the field of image dehazing, and proposed a dehazing algorithm for natural and optical remote sensing images based on BL-G-BM3D decomposition model. The proposed dehazing algorithm first utilized an adaptive histogram equalization to adjust the grayscale distribution of the image. Then the BL-G-BM3D decomposition model was applied for further enhancement, achieving effective dehazing. The proposed dehazing algorithm not only preserves image details effectively but also produces dehazed images that appear more natural and realistic, with better retention of details and edge information. Experimental results demonstrate that our proposed dehazing algorithm excels in both subjective and objective quality assessments, effectively handling natural and optical remote sensing images.