A Cross-View Matching Method Based on Dense Partition Strategy for UAV Geolocalization

This paper reports our solution for ACM Multimedia 2023 cross-view geo-localization challenge [29], which aims to solve real-world geo-localization task with extremely large satellite-view gallery distractors. Our solution is built on the basis of SwinV2 [13] and LPN [22]. Concretely, we adopt the SwinV2-B [13], the current mainstream transformer-based feature extractor, as the backbone of our model. Inspired by the feature partition strategy of LPN [22], we design a more efficient partition strategy named dense partition strategy. It segments and combines features to alleviate the problem of feature discontinuity at the boundary of different partitions. We get eight place in the geographic localization track on the official leaderboard.