A Deep Unfolded Prior-Aided RPCA Network for Cloud Removal

Clouds, together with their shadows, usually occlude ground-cover features in optical remote sensing images. This hinders the utilization of these images for a range applications such as earth observation, land-cover classification and urban planning. In this work, we propose deep unfolded prior-aided robust principal component analysis (DUPA-RPCA) network removing clouds recovering information multi-temporal satellite We model cloud-contaminated sum low rank sparse elements then unfold an iterative RPCA algorithm that has been designed reweighted $\ell _{1}$ minimization. As result, activation function DUPA-RPCA adapts every input at each layer network. Our experimental results on both Landsat Sentinel indicate our method gives better accuracy efficiency when compared existing state art methods.