ATTENTION/WARNING - NE PAS DÉPOSER ICI/DO NOT SUBMIT HERE

Ceci est la version de TEST de DIAL.mem. Veuillez ne pas soumettre votre mémoire sur ce site mais bien à l'URL suivante: 'https://thesis.dial.uclouvain.be'.
This is the TEST version of DIAL.mem. Please use the following URL to submit your master thesis: 'https://thesis.dial.uclouvain.be'.
 

Detection and tracking of railways in satellite images using deep learning

Lauwers, Adrien ; Samelson, Jonathan
(2020)

Files

Lauwers_60491700_Samelson_47301700_2020.pdf
  • Open access
  • Adobe PDF
  • 10.53 MB
Download

Details

Supervisors
Schaus, Pierre
Faculty
Ecole polytechnique de Louvain
Degree label
Master [120] en sciences informatiques
Abstract
Having an accurate and up-to-date network is essential for any transport service. This is particularly important in case of railway simulations where every distance has to be respected. At the moment, creating such a network needs human intervention and can take several days to reproduce a single railway line. Our work focuses on making the job easier by detecting the railroads in satellite images, using deep learning. In the recent years, deep neural networks have prevailed in image recognition tasks. This has been made possible thanks to the exponential exploding amount of data, the rise of computer power and the arrival of new algorithms. At the same time, the quality of satellite images has improved, allowing us to get enough details to enable such detections in most parts of the world. After the comparison of several deep convolutional neural network architectures, we build strong models which classify satellite images to detect the presence of railroads. We then use this underlying task to follow a railway line in order to reproduce its route as precisely as possible.

Citations

  • APA

Lauwers, A., & Samelson, J. (2020). Detection and tracking of railways in satellite images using deep learning. https://dial-mem.test.bib.ucl.ac.be/handle/2078.2/17523