Semi- and weak-supervised learning for Norwegian tree species detection
Keywords:deep learning, semi-supervised, semantic segmentation, tree species, remote sensing, aerial imagery
Tree species mapping of Norwegian production forests is a time-consuming process as forest associations largely rely on manual interpretation of earth observation data. Deep learning based image segmentation techniques have the potential to improve automated tree species classification, but a major challenge is the limited quality and availability of training data. Semi-supervised techniques could alleviate the need for training label and weak supervision enables handling coarse-grained and noisy labels. In this study, we evaluated the added value of semi-supervised deep learning methods in a weakly supervised setting. Specifically, consistency training and pseudo-labeling are applied for tree species classification from aerial ortho imagery in Norway. The techniques are generic and relevant for the wider earth observation domain, especially for other land cover segmentation tasks. The results show that consistency training gives a significant performance increase. Pseudo-labeling on the other hand does not, potentially this is due to varying convergence speeds for different classes causing confirmation bias or a partial violation of the cluster assumption.
Norge i bilder. URL https://norgeibilder.no/
J. Breidenbach, L. T. Waser, M. Debella-Gilo, J. Schumacher, J. Rahlf, M. Hauglin, S. Puliti, and R. Astrup. National mapping and estimation of forest area by dominant tree species using sentinel-2 data. Canadian Journal of Forest Research, 51(3):365–379, 2021. doi: 10.1139/cjfr-2020-0170.
M. Hauglin, J. Rahlf, J. Schumacher, R. Astrup, and J. Breidenbach. Large scale mapping of forest attributes using heterogeneous sets of airborne laser scanning and national forest inventory data. Forest Ecosystems, 8:1–15, 2021.
T. Hoeser and C. Kuenzer. Object detection and image segmentation with deep learning on earth observation data: A review-part i: Evolution and recent trends. Remote Sensing, 12 (10):1667, 2020. doi: 10.3390/rs12101667.
T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár. Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision, pages 2980–2988, 2017. doi: 10.1109/TPAMI.2018.2858826.
S. Nezami, E. Khoramshahi, O. Nevalainen, I. Pölönen, and E. Honkavaara. Tree species classification of drone hyperspectral and rgb imagery with deep learning convolutional neural networks. Remote Sensing, 12(7):1070, 2020. doi: 10.20944/preprints202002.0334.v1.
Y. Ouali, C. Hudelot, and M. Tami. An overview of deep semi-supervised learning. arXiv preprint arXiv:2006.05278, 2020. URL https://arxiv.org/pdf/2006.05278.pdf.
O. Ronneberger, P. Fischer, and T. Brox. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention, pages 234–241. Springer, 2015. doi: 10.1007/978-3-319-24574-4 28.
Q. Xie, Z. Dai, E. Hovy, T. Luong, and Q. Le. Unsupervised data augmentation for consistency training. Advances in Neural Information Processing Systems, 33:6256–6268, 2020. URL https://arxiv.org/abs/1904.12848.
Z.-H. Zhou. A brief introduction to weakly supervised learning. National science review, 5 (1):44–53, 2018. doi: 10.1093/NSR/NWX106.
Copyright (c) 2023 Martijn Vermeer, David Völgyes, Tord Kriznik Sørensen, Heidrun Miller, Daniele Fantin
This work is licensed under a Creative Commons Attribution 4.0 International License.