NSL-SLAM: High-Fidelity Neural Structured-Light Depth for Practical SLAM and Reconstruction

2026-07-27

AI Summary

This work addresses the challenge of achieving high-precision depth estimation, stable tracking, and high-quality reconstruction simultaneously in geometrically degenerate scenes, where existing SLAM systems often struggle. The authors propose a depth-centric, practical SLAM framework that, for the first time, integrates strong monocular depth priors into neural structured light decoding to enhance depth accuracy. The system adopts a depth-driven pipeline for both tracking and reconstruction, selectively incorporating sparse visual features and lightweight bundle adjustment only when necessary, thereby jointly optimizing robustness and efficiency. Experimental results demonstrate that the method reduces depth RMSE by 35% and improves reconstruction F-score by 1.6 points on Replica-SL, while avoiding catastrophic failures across all eight real-world sequences, achieving a 43.3% reduction in trajectory error and running online at 20.9 FPS.

Abstract

Structured-light (SL) cameras power depth sensing in millions of devices, and recent neural SL decoding methods have substantially improved their depth quality. SLAM systems can benefit greatly from such strong depth sensing, where reliable geometry enables stable tracking and faithful reconstruction. In this work, we present NSL-SLAM, a practical SLAM system tailored for high-fidelity structured-light depth. We first strengthen SL depth sensing: inspired by the neural structured-light (NSL) method, we further incorporate strong monocular depth priors into the SL stereo decoding, reducing depth RMSE by 35% on Replica-SL compared to NSL. We then build a depth-centric SLAM pipeline with this stronger depth: because structured-light geometry is dense and metrically accurate, we keep it as the primary tracking signal, and add only sparse visual correspondences for geometrically degenerate cases and lightweight bundle adjustment for long-range drift. Our depth estimator and SLAM design reinforce each other: stronger depth makes a simple SLAM pipeline effective, and the depth-centric pipeline ensures this advantage transfers to downstream reconstruction. Experimentally, on the synthetic Replica-SL benchmark, NSL-SLAM achieves the best tracking accuracy and improves reconstruction F-score by 1.6 points over the SOTA baseline under a shared-depth protocol. On a real benchmark of 8 challenging scenes, it is the only method that avoids catastrophic failure on all sequences while achieving 43.3% lower trajectory deviation than selected baselines. The SLAM system runs online at 20.9 FPS, demonstrating that stronger structured-light depth and depth-centric system design together enable practical, robust SLAM.

Problem

• Structured-light
• SLAM
• Depth sensing
• 3D reconstruction
• Robust tracking

Innovation

• Neural Structured-Light
• Depth-Centric SLAM
• Monocular Depth Prior
• High-Fidelity Depth
• Real-Time Reconstruction