Search papers, labs, and topics across Lattice.
WaterSplat-SLAM is introduced, a monocular SLAM system tailored for underwater environments, addressing the challenge of photorealistic rendering in this domain. It incorporates semantic medium filtering into two-view 3D reconstruction to improve camera tracking and depth estimation. The system also employs a semantic-guided rendering and adaptive map management strategy using an online medium-aware Gaussian map, resulting in photorealistic and compact underwater environment modeling.
Finally, underwater SLAM can produce photorealistic maps thanks to a novel medium-aware Gaussian map representation.
Underwater monocular SLAM is a challenging problem with applications from autonomous underwater vehicles to marine archaeology. However, existing underwater SLAM methods struggle to produce maps with high-fidelity rendering. In this letter, we propose WaterSplat-SLAM, a novel monocular underwater SLAM system that achieves robust pose estimation and photorealistic dense mapping. Specifically, we couple semantic medium filtering into two-view 3D reconstruction prior to enable underwater-adapted camera tracking and depth estimation. Furthermore, we present a semantic-guided rendering and adaptive map management strategy with an online medium-aware Gaussian map, modeling underwater environment in a photorealistic and compact manner. Experiments on multiple underwater datasets demonstrate that WaterSplat-SLAM achieves robust camera tracking and high-fidelity rendering in underwater environments.
Tsinghua AI