Cloud removal from satellite imagery using multispectral edge-filtered conditional generative adversarial networks

We propose a Generative Adversarial Network (GAN) based architecture for removing clouds from satellite imagery. Data used training comprises of visible light RGB and near-infrared (NIR) band images. The novelty lies in the structure discriminator GAN architecture, which compares generated target cloud-free images concatenated with their edge-filtered versions. Experimental results show that our approach to outperforms both visually according metrics, benchmark solution does not take edge filtering into account, improvements are robust when varying dataset size NIR cloud penetrability.