Simultaneous Geometrical Autofocusing and Multi-Image Alignment for UAV-Based SAR Systems

UAV-borne SAR technology is gaining traction in remote sensing due to its flexibility, rapid deployment, and cost-effectiveness. However, achieving high-quality SAR images requires precise platform localization within a small fraction of the wavelength, which remains challenging for low-cost UAVs. To mitigate this issue, autofocus algorithms are widely employed to compensate for trajectory deviations. This paper introduces a novel Geometrical Autofocus (GA) method that simultaneously performs autofocus and aligns multiple SAR images acquired from different UAV trajectories, serving as an initial co-registration step for InSAR applications. The GA method employs a geometrical approach to estimate the true UAV trajectory while fully accounting for the three-dimensional nature of the acquisition. It is applicable to all SAR imaging modes. The proposed method is validated using real UAV SAR data, demonstrating its effectiveness in correcting trajectory errors even in the absence of calibration corner reflectors. Additionally, two SAR images acquired from misaligned trajectories were successfully aligned, as evidenced by the improved interferogram coherence after the proposed calibration procedure.