SEN1-2 Dataset for Deep Learning in SAR-Optical Data Fusion published
In accordance with their terms of reference (SAR big data provided by Sentinel-1, fusion of SAR and complementary data), ISPRS WG I/3 is happy to announce the publication of the SEN1-2 dataset for the fostering of deep learning approaches in SAR-optical data fusion. SEN1-2 consists of 282,384 pairs of corresponding Sentinel-1 SAR and Sentinel-2 optical image patches, collected from across the globe and throughout all meteorological seasons. It is intended to support the introduction of computer vision approaches to multi-sensor remote sensing and can be used for studies on image matching, image-to-image translation, or image colorization, to name just a few examples.
The dataset was officially released in the frame of the ISPRS TCI Mid-Term Symposium 2018 in Karlsruhe with a corresponding paper published in the ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Interested colleagues can have a look at the paper at https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/IV-1/141/2018/.
The dataset can be downloaded from https://mediatum.ub.tum.de/1436631 under a CC-BY license.