It is a great pleasure to announce that our new dataset for the detection of sudden landslides following the 2018 earthquake in Haiti is now available online on IEEE Dataport.


The dataset is composed of 1713 samples, each containing a pre- and post-event optical images from Sentinel-2 associated with pre- and post-event SAR images from Sentinel-1. Each sample is assigned a ground truth label which was computed by NASA. The main objective of the dataset is to encourage the developement of multimodal approaches to exploit a post-event SAR image in place of an optical image for early response to natural disasters, here landslides.


This dataset was first introduced at the French Workshop GDR IASIS in Paris in the session named "Détection de changement dans les données géospatiales"


If our work can be of any use for use, please consider citing it as:


BRALET Antoine, TROUVÉ Emmanuel, CHANUSSOT Jocelyn, ATTO Abdourrahmane M., “Multimodal Remote Sensing Dataset for Landslide Change Detection in Haiti”, IEEE Dataport, July 14, 2024, doi:10.21227/4heb-7h07