CurriculumLoc: Enhancing Cross-Domain Geolocalization Through Multistage Refinement

• Published in: IEEE transactions on geoscience and remote sensing 2024, Vol.62, p.1-14
• Main Authors: Hu, Boni, Chen, Lin, Chen, Runjian, Bu, Shuhui, Han, Pengcheng, Li, Haowei
• Format: Article
• Language: English
• Subjects: Cross-domain geolocalization; Curriculum; Datasets; Design; Feature extraction; Geology; geometric verification; Location awareness; multistage geolocation refinement; Pipeline design; Seasonal variation; Seasonal variations; semantic attention; Semantics; Training; Transformers; visual localization; Visualization
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/TGRS.2024.3380191

Visual geolocalization is a cost-effective and scalable task that involves matching one or more query images, taken at some unknown location, to a set of geotagged reference images. Existing methods, devoted to semantic features representation, evolving toward robustness to a wide variety between query and reference, including illumination and viewpoint changes, as well as scale and seasonal variations. However, practical visual geolocalization approaches need to be robust in appearance changing and extreme viewpoint variation conditions, while providing accurate global location estimates. Therefore, inspired by curriculum design, human learn general knowledge first and then delve into professional expertise. We first recognize semantic scene and then measure geometric structure. Our approach, termed CurriculumLoc, involves a delicate design of multistage refinement pipeline and a novel keypoint detection and description with global semantic awareness and local geometric verification. We rerank candidates and solve a particular cross-domain perspective-n-point (PnP) problem based on these keypoints and corresponding descriptors, position refinement occurs incrementally. The extensive experimental results on our collected dataset, TerraTrack and a benchmark dataset, ALTO, demonstrate that our approach results in the aforementioned desirable characteristics of a practical visual geolocalization solution. Additionally, we achieve new high recall@1 scores of 62.6% and 94.5% on ALTO, with two different distances metrics, respectively. Dataset, code, and trained models are publicly available on https://github.com/npupilab/CurriculumLoc .