DAGS-SLAM: Dynamic-Aware 3DGS SLAM via Spatiotemporal Motion Probability and Uncertainty-Aware Scheduling
🤖 AI Summary
This work addresses the significant degradation in accuracy of 3D Gaussian Splatting (3DGS)-based SLAM systems on resource-constrained mobile devices caused by dynamic objects and occlusions. To tackle this challenge, we propose DAGS-SLAM, which innovatively models spatiotemporal motion probabilities for each Gaussian element and introduces an uncertainty-aware scheduler that triggers lightweight semantic analysis only when necessary. The system fuses YOLO instance priors with geometric cues to estimate motion probabilities, enabling the frontend to discard dynamic correspondences and the backend to suppress dynamic artifacts through probabilistic guidance during optimization. Experiments demonstrate that DAGS-SLAM achieves superior reconstruction quality and tracking robustness on public dynamic RGB-D datasets while maintaining real-time performance on commodity GPUs, confirming its efficiency and practicality for mobile deployment.
Application Category
📝 Abstract
Mobile robots and IoT devices demand real-time localization and dense reconstruction under tight compute and energy budgets. While 3D Gaussian Splatting (3DGS) enables efficient dense SLAM, dynamic objects and occlusions still degrade tracking and mapping. Existing dynamic 3DGS-SLAM often relies on heavy optical flow and per-frame segmentation, which is costly for mobile deployment and brittle under challenging illumination. We present DAGS-SLAM, a dynamic-aware 3DGS-SLAM system that maintains a spatiotemporal motion probability (MP) state per Gaussian and triggers semantics on demand via an uncertainty-aware scheduler. DAGS-SLAM fuses lightweight YOLO instance priors with geometric cues to estimate and temporally update MP, propagates MP to the front-end for dynamic-aware correspondence selection, and suppresses dynamic artifacts in the back-end via MP-guided optimization. Experiments on public dynamic RGB-D benchmarks show improved reconstruction and robust tracking while sustaining real-time throughput on a commodity GPU, demonstrating a practical speed-accuracy tradeoff with reduced semantic invocations toward mobile deployment.
Problem
Research questions and friction points this paper is trying to address.
Dynamic SLAM
3D Gaussian Splatting
Real-time Localization
Mobile Robotics
Occlusion Handling
Innovation
Methods, ideas, or system contributions that make the work stand out.
3D Gaussian Splatting
Dynamic SLAM
Motion Probability
Uncertainty-Aware Scheduling
Efficient Semantic Integration
🔎 Similar Papers
No similar papers found.
