Interferometric Processing of Radar Data Acquired by UAVs

Kurzfassung

Earth observation and remote sensing have made great advancements since the introduction of Synthetic Aperture Radar (SAR). From catastrophe monitoring and ship detection to deformation monitoring, SAR provides a lot of useful information at all-weather and all-day conditions. Unmanned Aerial Vehicle Synthetic Aperture Radar (UAV-SAR) has emerged in remote sensing offering significant advantages over traditional airborne and space-borne SAR systems. UAV-SAR systems offer distinct advantages such as flexibility, low power requirements, and high spatial resolutions. However, these advantages come with additional challenges and processing. One application of UAV-SAR is the formation of very high-quality digital elevation models (DEMs) using repeat-pass interferometry. Correcting phase errors caused by non-linear trajectories of UAVs is a major challenge. In addition, a common problem in interferometry is the phase unwrapping and the resulting phase errors. This research attempts to estimate both the motion errors and phase unwrapping errors and correct them. Acquisitions in multiple frequencies, multiple baselines, and radargrammetry are used to accomplish this. The best results are given by the combination of multiple frequencies and radargrammetry. However, the residual motion errors cannot be estimated separately from the phase unwrapping errors which influences the performance of the phase unwrapping detection. Furthermore, the degradation in the image focusing performance due to the use of efficient frequency-domain algorithms (omega-k) in the context of highly non-linear trajectories has been studied. The degradation of the image focusing performance is negligible for flat and steep terrain types as long as the necessary pre-processing is implemented.