DORF-EASNet: physics-driven real-time seafloor classification via entropy‑regularized acoustic features and adaptive model activation

Key Points

• Effective seafloor classification achieved through real-time processing with entropy-regularized acoustic features, improving accuracy.
• The system incorporates adaptive model activation, optimizing performance based on environmental conditions.
• A novel approach utilizing physics-driven principles enhances traditional seafloor classification methods.
• Potential applications in underwater exploration highlight the significance of accurate seafloor mapping for environmental monitoring.