Image-based Geolocalization by Ground-to-2.5D Map Matching
This addresses the problem of accurate geolocalization for applications like navigation and mapping, though it is incremental by extending existing cross-view matching with 2.5D data.
The paper tackles image-based geolocalization by matching ground-view images to 2.5D maps instead of 2D maps, using geometric height information to reduce cross-view appearance differences, and achieves significantly higher localization accuracy and faster convergence than previous methods.
We study the image-based geolocalization problem, aiming to localize ground-view query images on cartographic maps. Current methods often utilize cross-view localization techniques to match ground-view query images with 2D maps. However, the performance of these methods is unsatisfactory due to significant cross-view appearance differences. In this paper, we lift cross-view matching to a 2.5D space, where heights of structures (e.g., trees and buildings) provide geometric information to guide the cross-view matching. We propose a new approach to learning representative embeddings from multi-modal data. Specifically, we establish a projection relationship between 2.5D space and 2D aerial-view space. The projection is further used to combine multi-modal features from the 2.5D and 2D maps using an effective pixel-to-point fusion method. By encoding crucial geometric cues, our method learns discriminative location embeddings for matching panoramic images and maps. Additionally, we construct the first large-scale ground-to-2.5D map geolocalization dataset to validate our method and facilitate future research. Both single-image based and route based localization experiments are conducted to test our method. Extensive experiments demonstrate that the proposed method achieves significantly higher localization accuracy and faster convergence than previous 2D map-based approaches.