GeoFlow-SLAM++: A Robust Multi-Camera Visual-Inertial SLAM System with Relocalization

2026-06-20

AI Summary

This work addresses the limitations of monocular and RGB-D visual-inertial SLAM systems—such as restricted field of view, sensor failure susceptibility, and unreliable cross-session relocalization—by proposing a tightly coupled multi-camera visual-inertial SLAM framework. The system employs a body-centric modeling approach that integrates multi-camera reprojection constraints, IMU preintegration, and cross-view place recognition. It further enhances tracking robustness by combining dual-stream optical flow with learned neural features (SuperPoint/LightGlue) and introduces cross-view consistent pseudo-depth as an auxiliary geometric constraint. Supporting interchangeable frontends based on either ORB or neural features, extensive evaluations on EuRoC, TUM, OpenLORIS, Hilti, and custom datasets demonstrate that the multi-camera architecture significantly improves localization accuracy, neural features enhance robustness to appearance changes, and relocalization performance rivals that of LiDAR-based methods.

Abstract

Monocular and RGB-D visual-inertial SLAM systems remain susceptible to limited field of view, sensor-specific failure modes, and unreliable cross-session relocalization. To address these issues, we present GeoFlow-SLAM++, a tightly coupled multi-camera visual-inertial SLAM system that extends GeoFlow-SLAM from a single RGB-D sensor to a calibrated multi-camera rig with a unified body-centric formulation. Within this multi-camera framework, GeoFlow-SLAM++ supports two interchangeable visual front-ends: a conventional ORB front-end and a neural network feature (NN-Feature) front-end built on SuperPoint and LightGlue. The system unifies tracking, mapping, and relocalization on a shared body state, and combines multi-camera reprojection constraints, IMU pre-integration, cross-view place recognition, and dual-stream optical flow/NN-Feature tracking for robust localization. As an optional extension, the system can further incorporate cross-view-consistent pseudo-depth predictions from RGB images as auxiliary geometric constraints. We evaluate GeoFlow-SLAM++ on EuRoC, OpenLORIS, TUM, Hilti, and a self-collected handheld multi-camera dataset. Results show that the NN-Feature front-end improves robustness in appearance-challenging scenarios, the multi-camera formulation achieves competitive localization accuracy on Hilti, and the unified cross-view relocalization design reaches LiDAR-comparable performance on the handheld dataset.

Problem

• visual-inertial SLAM
• multi-camera
• relocalization
• field of view
• sensor failure

Innovation

• multi-camera SLAM
• visual-inertial odometry
• neural network features
• cross-view relocalization
• unified body-centric formulation