Detection of forest roads in Sentinel-2 images using U-Net


  • Øivind Trier Norsk Regnesentral
  • Arnt-Børre Salberg Norsk Regnesentral
  • Ragnvald Larsen Miljødirektoratet
  • Ole Torbjørn Nyvoll Miljødirektoratet



Remote sensing, Deep learning, Undisturbed nature, Nature interventions


This paper presents a new method for semi-automatic detection of nature interventions in
Sentinel-2 satellite images with 10 m spatial resolution. The Norwegian Environment Agency is
maintaining a map of undisturbed nature in Norway. U-Net was used for automated detection of
new roads, as these are often the cause whenever the area of undisturbed nature is reduced. The
method was able to detect many new roads, but with some false positives and possibly some false
negatives (i.e., missing new roads). In conclusion, we have demonstrated that automated detection of
new roads, for the purpose of updating the map of undisturbed nature, is possible. We have also
suggested several improvements of the method to improve its usefulness.


D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.

Q. Liu, M. Kampffmeyer, R. Jenssen, and A.B. Salberg. Road mapping in lidar images using a joint-task dense dilated convolutions merging network. In IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium, pages 5041{5044. IEEE, 2019. DOI:10.1109/IGARSS.2019.8900082.

Q. Liu, M. C. Kampffmeyer, R. Jenssen, and A.-B. Salberg. Multi-view self-constructing graph convolutional networks with adaptive class weighting loss for semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 44{45, 2020. DOI:10.1109/CVPRW50498.2020.00030.

A. D. Nobre, L. A. Cuartas, M. Hodnett, C. D. Renno, G. Rodrigues, A. Silveira, and S. Saleska. Height above the nearest drainage{ a hydrologically relevant new terrain model. Journal of Hydrology, 404(1-2):13{29, 2011. DOI: 10.1016/j.jhydrol.2011.03.051.

O. Ronneberger, P. Fischer, and T. Brox. U-Net: convolutional networks for biomedical image segmentation. In N. Navab, J. Hornegger, W. Wells, and A. Frangi, editors, Medical Image Computing and Computer-Assisted Intervention { MICCAI 2015. Part III. Lecture Notes in Computer Science, vol 9351, pages 234{241. Springer, 2017. DOI: 10.1007/978-3-319-24574-4 28.

A. B. Salberg, . D. Trier, and M. C. Kampffmeyer. Large-scale mapping of small roads in lidar images using deep convolutional neural networks. In Image Analysis 20th Scandinavian Conference, SCIA 2017 Tromso, Norway, June 12{14, 2017 Proceedings, Part II, page 193{204. Springer, 2017. DOI: 10.1007/978-3-319-59129-2 17.

S. Shit, J. C. Paetzold, A. Sekuboyina, I. Ezhov, A. Unger, A. Zhylka, J. P. Pluim, U. Bauer, and B. H. Menze. clDice-a novel topologypreserving loss function for tubular structure segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16560{16569,2021.

X. Yang, X. Li, Y. Ye, R. Y. K. Lau, X. Zhang, and X. Huang. Road detection and centerline extraction via deep recurrent convolutional neural network U-Net. IEEE Transactions on Geoscience and Remote Sensing, 57(9):7209{7220, 2019. DOI: 10.1109/TGRS.2019.2912301.