Flying Co-Stereo: Enabling Long-Range Aerial Dense Mapping via Collaborative Stereo Vision of Dynamic-Baseline

2025-05-31

AI Summary

To address the limited long-range dense mapping capability of UAV swarms in large-scale unknown environments, this paper proposes a lightweight long-distance dense mapping method based on collaborative stereo vision with dynamic inter-UAV baselines. We innovatively design a cross-UAV dual-spectrum (visible-infrared) baseline estimation algorithm to robustly estimate dynamic baselines; develop a hybrid feature association strategy integrating deep learning-based inter-UAV feature matching and onboard optical flow tracking; and introduce a sparse-to-dense depth recovery framework that enhances metric accuracy via exponential depth fitting of long-baseline triangulated point clouds. Experiments demonstrate that the system achieves dense 3D mapping up to 70 m, with relative depth errors of 2.3%–9.7%. Compared to conventional fixed short-baseline stereo systems, the proposed method improves maximum depth range and coverage area by 350% and 450%, respectively.

Technology Category

• Intelligent Robots: State Estimation
• Computer Vision: Low Level & Physics-based Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Lightweight long-range mapping is critical for safe navigation of UAV swarms in large-scale unknown environments. Traditional stereo vision systems with fixed short baselines face limited perception ranges. To address this, we propose Flying Co-Stereo, a cross-agent collaborative stereo vision system that leverages the wide-baseline spatial configuration of two UAVs for long-range dense mapping. Key innovations include: (1) a dual-spectrum visual-inertial-ranging estimator for robust baseline estimation; (2) a hybrid feature association strategy combining deep learning-based cross-agent matching and optical-flow-based intra-agent tracking; (3) A sparse-to-dense depth recovery scheme,refining dense monocular depth predictions using exponential fitting of long-range triangulated sparse landmarks for precise metric-scale mapping. Experiments demonstrate the Flying Co-Stereo system achieves dense 3D mapping up to 70 meters with 2.3%-9.7% relative error, outperforming conventional systems by up to 350% in depth range and 450% in coverage area. The project webpage: https://xingxingzuo.github.io/flying_co_stereo

Problem

• Enabling long-range dense mapping for UAV swarms
• Overcoming fixed short baseline limitations in stereo vision
• Achieving precise metric-scale mapping via collaborative stereo

Innovation

• Dual-spectrum estimator for baseline estimation
• Hybrid feature association with deep learning
• Sparse-to-dense depth recovery via exponential fitting