PhysDNet: Physics-Guided Decomposition Network of Side-Scan Sonar Imagery

TL;DR

PhysDNet is a physics-guided neural network that decomposes side-scan sonar images into interpretable physical components, enhancing robustness and interpretability for seafloor mapping and underwater sensing.

Contribution

It introduces a novel physics-guided multi-branch network that decouples sonar images into reflectivity, elevation, and loss, enabling self-supervised training without ground-truth labels.

Findings

Decomposed representations preserve geological structures.

Captures physically consistent illumination and attenuation.

Produces reliable shadow maps.

Abstract

Side-scan sonar (SSS) imagery is widely used for seafloor mapping and underwater remote sensing, yet the measured intensity is strongly influenced by seabed reflectivity, terrain elevation, and acoustic path loss. This entanglement makes the imagery highly view-dependent and reduces the robustness of downstream analysis. In this letter, we present PhysDNet, a physics-guided multi-branch network that decouples SSS images into three interpretable fields: seabed reflectivity, terrain elevation, and propagation loss. By embedding the Lambertian reflection model, PhysDNet reconstructs sonar intensity from these components, enabling self-supervised training without ground-truth annotations. Experiments show that the decomposed representations preserve stable geological structures, capture physically consistent illumination and attenuation, and produce reliable shadow maps. These findings demonstrate that physics-guided decomposition provides a stable and interpretable domain for SSS analysis, improving both physical consistency and downstream tasks such as registration and shadow interpretation.