Deep learning based automatic grounding line delineation in DInSAR interferograms

• Main Author: Ramanath Tarekere, Sindhu
• Other Authors: Krieger, Lukas, Floricioiu, Dana
• Format: Other/Unknown Material
• Language: English
• Published: Zenodo 2024
• Subjects: Grounding lines; Antarctica; Differential SAR Interferometry; Antarc*
• Online Access: https://doi.org/10.5281/zenodo.10785613

This dataset contains a small subset of the AIS_cci GLL product, which covers several key glaciers and the corresponding HED-delineated grounding lines generated from our automatic delineation pipeline. A description of the attributes of the AIS_cci GLL product is provided in the Product User Guide . We do not indicate the split of the interferograms into training, validation and test sets as the complete AIS_cci dataset is not open-access. We also provide eight double difference interferograms at 100 m pixel size to demonstrate the generation of the features stack. Please note, eight samples are not sufficient to train the neural network to achieve the delineation capability described in our work. The "UUID" attribute in both GeoJSON files is an identifier that links the vector geometries to the interferogram TiFF files.