Multi-temporal and multimodal

One of the first LULC semantic segmentation dataset.

Overview

Fast and easy inference

Check out the first results and inference exploration.

Access results

Overview

Get an overview of MultiSenGE dataset.

View details »

Paper

Paper presented during ISPRS 2022 in Nice, France.

View details »

Data

MultiSenGE dataset is available on Zenodo.

View details »

MultiSenGE Dataset

MultiSenGE is a benchmark dataset for multi-temporal and multi-modal land use land cover mapping. MultiSenGE contains 8,157 multi-temporal patches of Sentinel-1 and Sentinel-2 imagery over the Grand-Est region in France. A newly and highly reworked ground reference product (OCSGE2-GEOGRANDEST) is available for each patch.