Style Alignment-Based Dynamic Observation Method for UAV-View Geo-Localization

The task of unmanned aerial vehicle (UAV)-view geo-localization is to estimate the localization of a query satellite/drone image by matching it against a reference dataset consisting of drone/satellite images. Though tremendous strides have been made in feature alignment between satellite and drone views, vast differences in both inter and intraclass due to changes in viewpoint, altitude, and lighting remain a huge challenge. In this article, a style alignment-based dynamic observation method for UAV-view geo-localization is proposed to meet the above challenges from two perspectives: visual style transformation and surrounding noise control. Specifically, we introduce a style alignment strategy (SAS) to transform the diverse visual style of drone-view images into a unified satellite images visual style. Then, a dynamic observation module is designed to evaluate the spatial distribution of images by mimicking human observation habits. It is featured by the hierarchical attention block (HAB) with a dual-square-ring stream structure, to reduce surrounding noise and geographical deformation. In addition, we propose a deconstruction loss (Dc loss) to push away features of different geo-tags and squeeze knowledge from unmatched images by correlation calculation. The experimental results demonstrate the SOTA performance of our model on benchmarked datasets. In particular, when compared to the prior art on University-1652, our results surpass the best of them (FSRA), while only requiring
$2\times $
fewer parameters. The code will be released at
https://github.com/Xcco1/SA_DOM
.