DenseSplat: Densifying Gaussian Splatting SLAM with Neural Radiance Prior

2025-02-13

AI Summary

Existing 3D Gaussian Splatting (3DGS)-based SLAM suffers from severe map incompleteness and impractical deployment under sparse-view settings, primarily due to its reliance on numerous keyframes. Method: This work introduces NeRF-based radiance field priors into the 3DGS-SLAM framework for the first time, enabling geometry-aware primitive initialization and dynamic pruning. It further integrates loop closure detection with pose graph optimization—including bundle adjustment (BA)—to ensure geometric consistency and real-time rendering capability using only a minimal number of keyframes. Results: Evaluated on multiple large-scale datasets, our method achieves significant improvements over state-of-the-art approaches in tracking accuracy, map completeness, and rendering efficiency. It demonstrates that NeRF priors effectively enhance the robustness and fidelity of 3DGS modeling under sparse input conditions, yielding dense, geometrically consistent, and real-time renderable maps.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: State Estimation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Gaussian SLAM systems excel in real-time rendering and fine-grained reconstruction compared to NeRF-based systems. However, their reliance on extensive keyframes is impractical for deployment in real-world robotic systems, which typically operate under sparse-view conditions that can result in substantial holes in the map. To address these challenges, we introduce DenseSplat, the first SLAM system that effectively combines the advantages of NeRF and 3DGS. DenseSplat utilizes sparse keyframes and NeRF priors for initializing primitives that densely populate maps and seamlessly fill gaps. It also implements geometry-aware primitive sampling and pruning strategies to manage granularity and enhance rendering efficiency. Moreover, DenseSplat integrates loop closure and bundle adjustment, significantly enhancing frame-to-frame tracking accuracy. Extensive experiments on multiple large-scale datasets demonstrate that DenseSplat achieves superior performance in tracking and mapping compared to current state-of-the-art methods.

Problem

• Combining NeRF and 3DGS in SLAM
• Addressing sparse-view mapping challenges
• Enhancing tracking and mapping accuracy

Innovation

• Combines NeRF and 3DGS advantages
• Uses sparse keyframes for dense maps
• Implements geometry-aware sampling strategies