Exploring Generalisability of Self-Distillation with No Labels for SAR-Based Vegetation Prediction
This work addresses generalizability challenges for remote sensing applications, though it is incremental in nature.
The study investigated the generalizability of self-distilled models for vegetation prediction using SAR data, finding that embedding space separation in S1GRD datasets led to higher errors in unfamiliar regions, while GSSIC datasets showed overlapping embeddings.
In this work we pre-train a DINO-ViT based model using two Synthetic Aperture Radar datasets (S1GRD or GSSIC) across three regions (China, Conus, Europe). We fine-tune the models on smaller labeled datasets to predict vegetation percentage, and empirically study the connection between the embedding space of the models and their ability to generalize across diverse geographic regions and to unseen data. For S1GRD, embedding spaces of different regions are clearly separated, while GSSIC's overlaps. Positional patterns remain during fine-tuning, and greater distances in embeddings often result in higher errors for unfamiliar regions. With this, our work increases our understanding of generalizability for self-supervised models applied to remote sensing.