Multimodal Remote-Sensing Change Interpretation through Saliency-Guided Feature Correspondence and Geometry-Aware Scene Reconstruction across Variable Viewpoints, Resolutions, and Imaging Conditions in Earth Observation
Journal of geospace science · 28.07.2026 · 6 (2) · 12-24
DOI: 10.71465/jgs63 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of geospace science, online erschienen am 28.07.2026. Verfasst von Bethany A. Webb. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The rapid proliferation of Earth observation satellites has generated vast repositories of multimodal remote-sensing data, offering unprecedented opportunities for large-scale environmental monitoring and urban development analysis. However, interpreting structural and semantic changes across disparate sensors, such as Synthetic Aperture Radar and optical imagery, remains profoundly challenging due to extreme variations in viewpoints, spatial resolutions, and temporal imaging conditions. This paper introduces a comprehensive framework designed to address these fundamental limitations through the integration of saliency-guided feature correspondence and geometry-aware scene reconstruction. By employing a novel feature alignment paradigm that prioritizes salient regions, the proposed methodology mitigates the catastrophic failures typically associated with cross-sensor matching under severe radiometric distortion. Concurrently, a geometry-aware reconstruction module normalizes variable viewing angles and resolutions, projecting disparate temporal inputs into a unified three-dimensional spatial representation. Through extensive theoretical analysis and rigorous experimental validation on complex heterogeneous datasets, the integrated system demonstrates substantial improvements in change interpretation accuracy, robustness against seasonal and illumination variations, and computational efficiency. The results indicate that coupling geometric constraints with salient semantic features effectively resolves the ambiguity inherent in uncalibrated, multimodal remote-sensing streams, establishing a new state-of-the-art methodology for robust Earth observation analytics.
Zitieren
APA
Webb, B. A. (2026). Multimodal Remote-Sensing Change Interpretation through Saliency-Guided Feature Correspondence and Geometry-Aware Scene Reconstruction across Variable Viewpoints, Resolutions, and Imaging Conditions in Earth Observation. Journal of geospace science, 6(2), 12-24. https://doi.org/10.71465/jgs63
Vancouver
Webb BA. Multimodal Remote-Sensing Change Interpretation through Saliency-Guided Feature Correspondence and Geometry-Aware Scene Reconstruction across Variable Viewpoints, Resolutions, and Imaging Conditions in Earth Observation. Journal of geospace science. 2026;6(2):12-24. doi:10.71465/jgs63
BibTeX
@article{Webb2026,
author = {Webb, Bethany A.},
title = {Multimodal Remote-Sensing Change Interpretation through Saliency-Guided Feature Correspondence and Geometry-Aware Scene Reconstruction across Variable Viewpoints, Resolutions, and Imaging Conditions in Earth Observation},
journal = {Journal of geospace science},
year = {2026},
volume = {6},
number = {2},
pages = {12--24},
doi = {10.71465/jgs63}
}
EUVIMED