Remote sensing image dehazing network integrating dynamic multi-expert gating

To address the issues of highly non-uniform fog distribution, severe color distortion, and easy loss of structural details in remote sensing images under complex atmospheric conditions, a remote sensing image dehazing network based on a dynamic multi-expert gating mechanism (dynamic multi-expert gating dehazing network, DMEGD-Net) is proposed on the basis of the U-Net framework. The file contains the model code and the weights trained on the SateHaze1k, HRSD and RRSHID datasets.