Semi-Supervised Learning for Joint SAR and Multispectral Land Cover Classification

Semi-supervised learning techniques are gaining popularity due to their capability of building models that effective, even when scarce amounts labeled data available. In this paper, we present a framework and specific tasks for self-supervised pretraining \textit{multichannel} models, such as the fusion multispectral synthetic aperture radar images. We show proposed approach is highly effective at features correlate with labels land cover classification. This enabled by an explicit design which promotes bridging gaps between sensing modalities exploiting spectral characteristics input. semi-supervised setting, limited available, using pretraining, followed supervised finetuning classification SAR data, outperforms conventional approaches purely learning, initialization from training on ImageNet other recent approaches.