Bernhard Rabus

Simon Fraser University, Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)

Papers

Total Citations

H-Index

About

Bernhard Rabus is a leading figure in radar remote sensing and deep learning for geospatial analysis. His research bridges the gap between synthetic aperture radar (SAR) technology and advanced artificial intelligence, with a primary focus on enhancing the reliability of automated systems. In his highly cited 2021 work (31 citations), Rabus tackled a critical challenge in mission-critical automation: the inability of deep learning models to predict their own failures. By developing robust uncertainty estimation methods for road segmentation in SAR imagery, he provided a pathway to safer, more trustworthy autonomous systems that can flag when human intervention is needed. Rabus also made foundational contributions to one of the most ambitious Earth observation missions ever undertaken. As a key contributor to the Shuttle Radar Topography Mission (SRTM), he co-authored the seminal 2002 paper (27 citations) detailing the X-band SAR products and processing facility. This work was instrumental in creating the first near-global, high-resolution digital elevation model, which remains a cornerstone dataset for geoscience, hydrology, and disaster management. Through his dual expertise in hardware-driven space missions and cutting-edge AI, Rabus continues to shape how we observe and interpret our planet from orbit.