arXiv 2605.103452026-05-11cs.CV

BGG: 通过基于视觉基础模型的适应方法弥合跨视角图像间的几何差距以实现地理定位

BGG: Bridging the Geometric Gap between Cross-View images by Vision Foundation Model Adaptation for Geo-Localization

Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China, Xidian University（中国教育部智能感知与图像理解重点实验室，西安电子科技大学）；School of Telecommunications, Xidian University（西安电子科技大学电信学院）；Department of Automation, Tsinghua University（清华大学自动化系）

Wei Wang, Dou Quan, Ning Huyan, Shuang Wang, Yi Li, Pei He, Licheng Jiao

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出BGG框架，通过多粒度特征增强适配器和频率感知结构聚合模块，有效提升跨视角地理定位性能，实验表明其在低训练成本下达到最先进的定位效果。

摘要：

跨视角图像间的几何差异，如无人机与卫星图像，显著增加了跨视角地理定位（CVGL）的挑战，旨在通过图像检索获取图像的地理定位。为进一步提升CVGL性能，本文提出一种参数高效的适应框架，基于视觉基础模型（VFM）（如DINOv3），称为BGG。BGG不仅有效利用了VFM的一般视觉表示，捕捉了跨视角图像的鲁棒且一致的特征，还利用了VFM的泛化能力，显著提升了CVGL性能。它主要包含多粒度特征增强适配器（MFEA）和频率感知结构聚合（FASA）模块。具体而言，MFEA通过多级扩张卷积增强特征的尺度适应性和视角鲁棒性，有效利用少量训练成本弥合跨视角几何差距。此外，考虑到[CLS]标记缺乏空间细节以实现精确的图像检索和定位，FASA模块在频域内调节补丁标记，并对局部结构特征进行自适应聚合。最后，BGG将增强的局部特征与[CLS]标记融合，以实现更准确的CVGL。在University-1652和SUES-200数据集上的广泛实验表明，BGG在其他方法上具有显著优势，并在低训练成本下实现了最先进的定位性能。

英文摘要：

Geometric differences between cross-view images, such as drone and satellite views, significantly increase the challenge of Cross-View Geo-Localization (CVGL), which aims to acquire the geolocation of images by image retrieval. To further enhance the CVGL performance, this paper proposes a parameter-efficient adaptation framework for bridging the geometric gap across images based on the vision foundation model (VFM) (e.g., DINOv3), termed BGG. BGG not only effectively leverages the general visual representations of VFM and captures the robust and consistent features from cross-view images, but also utilizes the generalization capabilities of the VFM, significantly improving the CVGL performance. It mainly contains a Multi-granularity Feature Enhancement Adapter (MFEA) and a Frequency-Aware Structural Aggregation (FASA) module. Specifically, MFEA enhances the scale adaptability and viewpoint robustness of features by multi-level dilated convolutions, effectively bridging the cross-view geometric gap with small training costs. Additionally, considering the [CLS] token lacks spatial details for precise image retrieval and localization, the FASA module modulates patch tokens in the frequency domain and performs adaptive aggregation for local structural feature enhancement. Finally, BGG fuses the enhanced local features with the [CLS] token for more accurate CVGL. Extensive experiments on University-1652 and SUES-200 datasets demonstrate that BGG has significant advantages over other methods and achieves state-of-the-art localization performance with low training costs.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。