AquaGS: Fast Underwater Scene Reconstruction with SfM-Free Gaussian Splatting

TL;DR

AquaGS is a rapid underwater scene reconstruction method that bypasses SfM, using Gaussian splatting and neural rendering to achieve high-precision 3D models from minimal images in under 30 seconds.

Contribution

The paper introduces AquaGS, a novel SfM-free underwater reconstruction approach combining Gaussian splatting and neural radiance fields for fast, accurate 3D modeling.

Findings

Reconstructs high-precision 3D models in 30 seconds

Requires only 3 images for effective reconstruction

Outperforms traditional SfM-based methods in speed and accuracy

Abstract

Underwater scene reconstruction is a critical tech-nology for underwater operations, enabling the generation of 3D models from images captured by underwater platforms. However, the quality of underwater images is often degraded due to medium interference, which limits the effectiveness of Structure-from-Motion (SfM) pose estimation, leading to subsequent reconstruction failures. Additionally, SfM methods typically operate at slower speeds, further hindering their applicability in real-time scenarios. In this paper, we introduce AquaGS, an SfM-free underwater scene reconstruction model based on the SeaThru algorithm, which facilitates rapid and accurate separation of scene details and medium features. Our approach initializes Gaussians by integrating state-of-the-art multi-view stereo (MVS) technology, employs implicit Neural Radiance Fields (NeRF) for rendering translucent media and utilizes the latest explicit 3D Gaussian Splatting (3DGS) technique to render object surfaces, which effectively addresses the limitations of traditional methods and accurately simulates underwater optical phenomena. Experimental results on the data set and the robot platform show that our model can complete high-precision reconstruction in 30 seconds with only 3 image inputs, significantly enhancing the practical application of the algorithm in robotic platforms.