August 15, 2025Remote SensingOpen Access

Deep Learning Small Water Body Mapping by Transfer Learning from Sentinel-2 to PlanetScope

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Transfer learning improves mapping accuracy of small water bodies in regional areas, indicating reduced manual annotation needs.

VMamba model improves small water body mapping accuracy, addressing constraints of manual annotations and data scarcity.
Achieved over 5648 patches tested in two separate areas of 9636 km2, showing enhanced precision compared to CNN and Transformer models.
Transfer learning employs pre-trained Sentinel-2 water masks to effectively fine-tune PlanetScope data.
Combining Sentinel-2 and PlanetScope resources may enable broader applications with fewer labor demands for water body mapping.

Li et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark