Underwater360: Reconstructing Underwater Scenes from Panoramic Images with Omnidirectional Gaussian Splatting

2026-05-25

AI Summary

Underwater scene reconstruction is hindered by medium-induced absorption and scattering effects, as well as the limited field of view of conventional cameras, leading to geometric distortions and appearance degradation. This work proposes a physics-guided omnidirectional 3D Gaussian splatting framework that performs ray casting directly in spherical camera space, explicitly decoupling intrinsic scene radiance from depth-dependent backscatter and attenuation. By conducting Gaussian splatting in spherical coordinates, the method mitigates projection distortions inherent in planar representations, while a pose-conditioned appearance embedding ensures physically consistent underwater appearance recovery. Evaluated on a newly introduced underwater panoramic dataset, the approach significantly improves novel-view synthesis quality and cross-view consistency, yielding more realistic and stable reconstructions.

Abstract

Underwater scene reconstruction is essential for immersive exploration of aquatic environments, yet remains challenging due to complex participating-media effects such as absorption and scattering, as well as the limited field of view (FoV) of conventional cameras. Although combining panoramic imaging with 3D Gaussian Splatting (3DGS) offers a promising direction for photorealistic underwater rendering, traditional 3DGS struggles with both spherical projection distortion and underwater medium degradation. In this paper, we propose \textbf{Underwater360}, a physics-informed omnidirectional 3DGS framework for underwater panoramic scene reconstruction. First, we introduce an Omnidirectional Gaussian Splatting module that performs ray casting directly in spherical camera space instead of relying on 2D projection approximations, thereby reducing geometric distortions under 360$^\circ$ FoV. Second, we design a physics-based appearance-medium modeling architecture with pose-conditioned appearance embeddings to explicitly decouple intrinsic scene radiance from depth-dependent backscatter and attenuation, enabling physically grounded scene appearance restoration. Finally, we establish a new panoramic underwater benchmark dataset containing both synthetic and real-world scenes. Extensive experiments demonstrate that Underwater360 achieves superior performance in underwater novel view synthesis and scene appearance restoration, delivering improved rendering quality and cross-view consistency in complex underwater environments. The code and datasets are released at https://github.com/SwcK423/Underwater360

Problem

• underwater scene reconstruction
• participating media effects
• panoramic imaging
• spherical projection distortion
• field of view

Innovation

• Omnidirectional Gaussian Splatting
• Underwater Scene Reconstruction
• Physics-based Rendering
• Panoramic Imaging
• Participating Media Effects