SpatialFly: Geometry-Guided Representation Alignment for UAV Vision-and-Language Navigation in Urban Environments

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jiang, Wen, Huang, Kangyao, Wang, Li, Xu, Wang, Fan, Wei, Liu, Jinyuan, Liu, Shaoyu, Liang, Hanfang, Duan, Hongwei, Xu, Bin, Ji, Xiangyang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910062786641920
author Jiang, Wen
Huang, Kangyao
Wang, Li
Xu, Wang
Fan, Wei
Liu, Jinyuan
Liu, Shaoyu
Liang, Hanfang
Duan, Hongwei
Xu, Bin
Ji, Xiangyang
author_facet Jiang, Wen
Huang, Kangyao
Wang, Li
Xu, Wang
Fan, Wei
Liu, Jinyuan
Liu, Shaoyu
Liang, Hanfang
Duan, Hongwei
Xu, Bin
Ji, Xiangyang
contents UAVs play an important role in applications such as autonomous exploration, disaster response, and infrastructure inspection. However, UAV VLN in complex 3D environments remains challenging. A key difficulty is the structural representation mismatch between 2D visual perception and the 3D trajectory decision space, which limits spatial reasoning. To this end, we propose SpatialFly, a geometry-guided spatial representation framework for UAV VLN. Operating on RGB observations without explicit 3D reconstruction, SpatialFly introduces a geometry-guided 2D representation alignment mechanism. Specifically, the geometric prior injection module injects global structural cues into 2D semantic tokens to provide scene-level geometric guidance. The geometry-aware reparameterization module then aligns 2D semantic tokens with 3D geometric tokens through cross-modal attention, followed by gated residual fusion to preserve semantic discrimination. Experimental results show that SpatialFly consistently outperforms state-of-the-art UAV VLN baselines across both seen and unseen environments, reducing NE by 4.03m and improving SR by 1.27% over the strongest baseline on the unseen Full split. Additional trajectory-level analysis shows that SpatialFly produces trajectories with better path alignment and smoother, more stable motion.
format Preprint
id arxiv_https___arxiv_org_abs_2603_21046
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SpatialFly: Geometry-Guided Representation Alignment for UAV Vision-and-Language Navigation in Urban Environments
Jiang, Wen
Huang, Kangyao
Wang, Li
Xu, Wang
Fan, Wei
Liu, Jinyuan
Liu, Shaoyu
Liang, Hanfang
Duan, Hongwei
Xu, Bin
Ji, Xiangyang
Computer Vision and Pattern Recognition
Artificial Intelligence
UAVs play an important role in applications such as autonomous exploration, disaster response, and infrastructure inspection. However, UAV VLN in complex 3D environments remains challenging. A key difficulty is the structural representation mismatch between 2D visual perception and the 3D trajectory decision space, which limits spatial reasoning. To this end, we propose SpatialFly, a geometry-guided spatial representation framework for UAV VLN. Operating on RGB observations without explicit 3D reconstruction, SpatialFly introduces a geometry-guided 2D representation alignment mechanism. Specifically, the geometric prior injection module injects global structural cues into 2D semantic tokens to provide scene-level geometric guidance. The geometry-aware reparameterization module then aligns 2D semantic tokens with 3D geometric tokens through cross-modal attention, followed by gated residual fusion to preserve semantic discrimination. Experimental results show that SpatialFly consistently outperforms state-of-the-art UAV VLN baselines across both seen and unseen environments, reducing NE by 4.03m and improving SR by 1.27% over the strongest baseline on the unseen Full split. Additional trajectory-level analysis shows that SpatialFly produces trajectories with better path alignment and smoother, more stable motion.
title SpatialFly: Geometry-Guided Representation Alignment for UAV Vision-and-Language Navigation in Urban Environments
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2603.21046