SP-SLAM: Neural Real-Time Dense SLAM With Scene Priors

2025-01-11

IEEE transactions on circuits and systems for video technology (Print)

AI Summary

Existing neural implicit SLAM methods suffer from ambiguous reconstructions and poor real-time performance, primarily due to ineffective modeling of scene priors. This paper introduces SP-SLAM—the first real-time RGB-D dense SLAM system integrating sparse voxel encoding priors with tri-plane representations. Its core contributions are: (1) a surface-proximity-aware sparse voxel prior that accelerates implicit field convergence; (2) inter-frame global voxel fusion coupled with joint online optimization over all historical poses; and (3) a lightweight tri-plane feature storage scheme that balances texture fidelity and computational efficiency under memory constraints. Evaluated on five standard benchmarks—including Replica—SP-SLAM achieves significant improvements in pose accuracy and reconstruction completeness while sustaining >30 FPS real-time performance, consistently outperforming state-of-the-art methods.

Technology Category

• Intelligent Robots: State Estimation
• Computer Vision: Representation Learning for Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Neural implicit representations have recently shown promising progress in dense Simultaneous Localization And Mapping (SLAM). However, existing works have shortcomings in terms of reconstruction quality and real-time performance, mainly due to inflexible scene representation strategy without leveraging any prior information. In this paper, we introduce SP-SLAM, a novel neural RGB-D SLAM system that performs tracking and mapping in real-time. SP-SLAM computes depth images and establishes sparse voxel-encoded scene priors near the surfaces to achieve rapid convergence of the model. Subsequently, the encoding voxels computed from single-frame depth image are fused into a global volume, which facilitates high-fidelity surface reconstruction. Simultaneously, we employ tri-planes to store scene appearance information, striking a balance between achieving high-quality geometric texture mapping and minimizing memory consumption. Furthermore, in SP-SLAM, we introduce an effective optimization strategy for mapping, allowing the system to continuously optimize the poses of all historical input frames during runtime without increasing computational overhead. We conduct extensive evaluations on five benchmark datasets (Replica, ScanNet, TUM RGB-D, Synthetic RGB-D, 7-Scenes). The results demonstrate that, compared to existing methods, we achieve superior tracking accuracy and reconstruction quality, while running at a significantly faster speed.

Problem

• Neural Implicit Representations
• SLAM
• Real-time Processing

Innovation

• Real-time Dense SLAM
• Depth Image Accelerated Learning
• Optimized Tracking and Reconstruction