Object-level Cross-view Geo-localization with Location Enhancement and Multi-Head Cross Attention

TL;DR

This paper introduces OCGNet, a novel object-level cross-view geo-localization network that leverages location cues, enhancement modules, and multi-head attention to improve accuracy and generalization in challenging scenarios.

Contribution

It proposes a new network architecture that integrates location information and attention mechanisms for precise object-level geo-localization across views.

Findings

Achieves state-of-the-art results on CVOGL dataset.

Demonstrates effective few-shot learning capabilities.

Robustly localizes objects despite viewpoint and condition variations.

Abstract

Cross-view geo-localization determines the location of a query image, captured by a drone or ground-based camera, by matching it to a geo-referenced satellite image. While traditional approaches focus on image-level localization, many applications, such as search-and-rescue, infrastructure inspection, and precision delivery, demand object-level accuracy. This enables users to prompt a specific object with a single click on a drone image to retrieve precise geo-tagged information of the object. However, variations in viewpoints, timing, and imaging conditions pose significant challenges, especially when identifying visually similar objects in extensive satellite imagery. To address these challenges, we propose an Object-level Cross-view Geo-localization Network (OCGNet). It integrates user-specified click locations using Gaussian Kernel Transfer (GKT) to preserve location information throughout the network. This cue is dually embedded into the feature encoder and feature matching blocks, ensuring robust object-specific localization. Additionally, OCGNet incorporates a Location Enhancement (LE) module and a Multi-Head Cross Attention (MHCA) module to adaptively emphasize object-specific features or expand focus to relevant contextual regions when necessary. OCGNet achieves state-of-the-art performance on a public dataset, CVOGL. It also demonstrates few-shot learning capabilities, effectively generalizing from limited examples, making it suitable for diverse applications (https://github.com/ZheyangH/OCGNet).