Sparse is Fast: Accelerating Synthetic Aperture Sonar via Reduced Sampling

Abstract

High-resolution underwater imaging is central to deep-sea exploration, marine science, and subsea infrastructure inspection. As interest in deep-sea missions grows - from ecological monitoring and resource management to defense and offshore energy - autonomous underwater vehicles (AUVs) have become key enablers. These mobile robots must traverse large, complex terrains where visibility is minimal, and GPS is unavailable. Acoustic imaging, particularly synthetic aperture sonar (SAS), is a leading technique for underwater scene reconstruction, offering high spatial resolution with compact form factors. However, its widespread deployment remains limited by one critical constraint: the speed at which imaging can be performed.