Vehicle Geolocalization from Drone Camera Imagery in Challenging Environments
Loading...
Date
2023-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
We have developed a robust and novel method for determining the geolocation of vehicles observed in drone camera-acquired images. Unlike previous works, which use reliable GPS positions and camera geometry to calculate the position of objects in drone footage, our method is robust to camera orientation and poor metadata for the drone's flight. The data we used is publicly available. Additionally, all the components and methods used are publicly available as well. Our method uses transformations between drone camera footage and Google Maps Aerial camera images to determine the positions of objects captured in the drone camera footage. We have explored different feature extraction methods, key point filtering schemes, and types of transformations. Our method is robust and only requires an estimated geolocation for a drone to determine the geolocation of vehicles on the ground.
Description
Keywords
drones, geolocalization, computer vision, gps, google maps