Satellite Image-based Localization via Learned Embeddings
This addresses the localization problem for autonomous vehicles or robotics in GPS-denied areas, representing a novel method for a known bottleneck.
The paper tackles the problem of localizing a ground vehicle using only satellite imagery as prior knowledge, by learning location-discriminative embeddings to match ground-level images with satellite views, and demonstrates successful localization in novel environments despite viewpoint and appearance variations.
We propose a vision-based method that localizes a ground vehicle using publicly available satellite imagery as the only prior knowledge of the environment. Our approach takes as input a sequence of ground-level images acquired by the vehicle as it navigates, and outputs an estimate of the vehicle's pose relative to a georeferenced satellite image. We overcome the significant viewpoint and appearance variations between the images through a neural multi-view model that learns location-discriminative embeddings in which ground-level images are matched with their corresponding satellite view of the scene. We use this learned function as an observation model in a filtering framework to maintain a distribution over the vehicle's pose. We evaluate our method on different benchmark datasets and demonstrate its ability localize ground-level images in environments novel relative to training, despite the challenges of significant viewpoint and appearance variations.