Deficit-Driven Stratified Acquisition for Building STRATA-S12: An Open SAR-Optical Corpus for Self-Supervised Learning
Updated
Updated · arxiv.org · Oct 22
Deficit-Driven Stratified Acquisition for Building STRATA-S12: An Open SAR-Optical Corpus for Self-Supervised Learning
1 articles · Updated · arxiv.org · Oct 22
Summary
Researchers have introduced DDSA, a new algorithm for building balanced multi-modal Earth observation datasets, releasing the STRATA-S12 SAR–optical corpus.
DDSA ensures even geographic and seasonal coverage by dynamically prioritising under-sampled regions, overcoming cloud-related data gaps in satellite imagery.
STRATA-S12, covering all inhabited continents, is designed for self-supervised learning and aims to improve the fairness and effectiveness of geospatial AI models.