Leveraging neural networks and edge detection for better UAV localization 🚁🗺️📍

Official implementation for the paper "Leveraging neural networks and edge detection for better UAV localization".

Paper accepted to IGARSS 2024 : arXiv submission

Method Overview

Getting Started

Clone the Repository

To clone this repository, use the following command:

git clone https://github.com/TheoDpPro/uav-localization.git

Installation

Make sure you have Python 3 installed. Then, install the dependencies using:

pip install -r requirements.txt

Training

To train the model, run the following command:

python train.py

Data

Example data is available in example-data folder. train.csv, test.csv contain filenames and coordinates of each tile. Below the structure of the data folder for n reference tiles and m uav views.

├── train/
│   ├── reference_tile_1.npy
│   ├── reference_tile_2.npy
│   ├── ...
│   └── reference_tile_n.npy
│
├── test/
│   ├── uav_view_1.npy
│   ├── uav_view_2.npy
│   ├── ...
│   └── uav_view_m.npy
│
├── train.csv
└── test.csv

Acknowledgment

Thanks to ABGRALL Corentin, BASCLE Benedicte, DAVAUX Jean-Clément, FACCIOLO Gabriele and MEINHARDT-LLOPIS Enric.

Citation

This project is based on the work by Di Piazza et al. If you use this code in your research, please cite the following paper:

@inproceedings{dipiazza2024uavloc,
  author    = {Di Piazza Theo, Meinhardt-Llopis Enric, Facciolo Gabriele, Bascle Benedicte, Abgrall Corentin and Devaux Jean-Clement},
  title     = {Leveraging neural networks and edge detection for better UAV localization},
  booktitle = {Proceedings of the IEEE International Geoscience and Remote Sensing Symposium (IGARSS)},
  year      = {2024},
  organization = {IEEE},
}