The study proposes using a visual base model to adapt acoustic signals, enabling 3D sonar reconstruction without the need for pose estimation.

On September 5, 2026, arXiv cs.CV published the paper “Adapting Vision Foundation Models to Acoustics for Pose-Free 3D Sonar Reconstruction”. This study proposes a method of adapting vision-based models to acoustic signals, aiming to achieve 3D sonar reconstruction without pose estimation.