EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Lightweight 3DGS-SLAM for Memory-Constrained Environments: Spatial-Aware Truncation and Adaptive Antihallucination Restoration Mechanism

Honghui Fan, Zikai Li, Hongjin Zhu, Wenhe Chen

ISPRS International Journal of Geo-Information · 06.07.2026 · 15 (7) · 306

DOI: 10.3390/ijgi15070306 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ISPRS International Journal of Geo-Information, online erschienen am 06.07.2026. Verfasst von Honghui Fan und 3 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, National Key Research and Development Program of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dense simultaneous localization and mapping (SLAM) via 3D Gaussian splatting (3DGS) faces memory bottlenecks due to the explosive growth of primitives during long-sequence mapping. We propose SATA-SLAM, a framework featuring spatial-aware truncation and adaptive anti-hallucination. The online front-end maintains a constant memory footprint via a spatial-aware pruning module (SAPM), which employs a survival scoring function that couples primitive opacity with view-frustum projection coverage and a temporal protection window. Subsequently, an anti-hallucination generative refinement module (AGRM) utilizes texture priors from pretrained diffusion models for offline inpainting of residual regions. In addition, an adaptive gating mechanism to verify and suppress AIGC-induced hallucinations caused by pose drift, ensuring multiview consistency. Experiments on the public Replica dataset show that SATA-SLAM improves rendering quality from 12.5 dB to 37.44 dB (averaged over the Replica room0 and office0 scenes) while using only 26% of the original memory, outperforming the unconstrained baseline. This study provides a pathway toward low-power, high-fidelity environmental perception for mobile robots.

Förderung

  • National Natural Science Foundation of China
  • National Key Research and Development Program of China

Zitieren

APA

Fan, H., Li, Z., Zhu, H., & Chen, W. (2026). Lightweight 3DGS-SLAM for Memory-Constrained Environments: Spatial-Aware Truncation and Adaptive Antihallucination Restoration Mechanism. ISPRS International Journal of Geo-Information, 15(7), 306. https://doi.org/10.3390/ijgi15070306

Vancouver

Fan H, Li Z, Zhu H, Chen W. Lightweight 3DGS-SLAM for Memory-Constrained Environments: Spatial-Aware Truncation and Adaptive Antihallucination Restoration Mechanism. ISPRS International Journal of Geo-Information. 2026;15(7):306. doi:10.3390/ijgi15070306

BibTeX
@article{Fan2026,
  author = {Fan, Honghui and Li, Zikai and Zhu, Hongjin and Chen, Wenhe},
  title = {Lightweight 3DGS-SLAM for Memory-Constrained Environments: Spatial-Aware Truncation and Adaptive Antihallucination Restoration Mechanism},
  journal = {ISPRS International Journal of Geo-Information},
  year = {2026},
  volume = {15},
  number = {7},
  pages = {306},
  doi = {10.3390/ijgi15070306}
}
Kostenlos starten →