SkyCap: Bitemporal VHR Optical-SAR Quartets for Amplitude Change Detection and Foundation-Model Evaluation

TL;DR

SkyCap introduces a novel dataset combining optical and SAR imagery for change detection, enabling evaluation of foundation models and revealing the importance of preprocessing alignment in SAR amplitude change detection.

Contribution

The paper presents SkyCap, the first VHR optical-SAR dataset for change detection, and benchmarks foundation models, highlighting the impact of preprocessing choices on SAR and optical model performance.

Findings

Optical foundation models outperform SAR-specific models on SAR ACD with proper preprocessing.

Pretraining on SAR data improves SAR model performance but is sensitive to preprocessing.

Model performance varies significantly with preprocessing alignment and data modality.

Abstract

Change detection for linear infrastructure monitoring requires reliable high-resolution data and regular acquisition cadence. Optical very-high-resolution (VHR) imagery is interpretable and straightforward to label, but clouds break this cadence. Synthetic Aperture Radar (SAR) enables all-weather acquisitions, yet is difficult to annotate. We introduce SkyCap, a bitemporal VHR optical-SAR dataset constructed by archive matching and co-registration of (optical) SkySat and Capella Space (SAR) scenes. We utilize optical-to-SAR label transfer to obtain SAR amplitude change detection (ACD) labels without requiring SAR-expert annotations. We perform continued pretraining of SARATR-X on our SAR data and benchmark the resulting SAR-specific foundation models (FMs) together with SARATR-X against optical FMs on SkyCap under different preprocessing choices. Among evaluated models, MTP(ViT-B+RVSA), an optical FM, with dB+Z-score preprocessing attains the best result (F1$_c$ = 45.06), outperforming SAR-specific FMs further pretrained directly on Capella data. We observe strong sensitivity to preprocessing alignment with pretraining statistics, and the ranking of optical models on optical change detection does not transfer one-to-one to SAR ACD. To our knowledge, this is the first evaluation of foundation models on VHR SAR ACD.