Noise immunity of the U-Net fully convolutional neural network model for semantic segmentation of fir trees on UAV images
DOI: 10.21293/1818-0442-2024-27-2-64-70
Abstract: The paper studies the noise immunity of a modified model of a fully convolutional neural network U-Net with robust Cauchy loss function and Focal Loss function when solving the prob-lem of segmentation (pixel classification) of noisy images of fir trees infected by pests. It is shown that the accuracy of mul-ticlassification of such trees decreases with the increase of area noise and the amplitude of impulse noise on the fragments of the training sample. At the same time, the level of accuracy degradation depends on the modified U-Net loss function used for training. For the model with robust Cauchy loss function, there is a slower decrease in noise immunity with increasing values of noise parameters.
Keywords: unmanned aerial vehicle, semantic segmentation of fir tree images, Noise immunity of a modified U-Net fully convolutional neural network model
For citation:
Malkin A. Yu., Markov N. G. Noise immunity of the U-Net fully convolutional neural network model for semantic segmentation of fir trees on UAV images. Doklady Tomskogo gosudarstvennogo universiteta sistem upravleniya i radioelektroniki, 2024, vol. 27, no. 2, pp. 64–70. DOI: 10.21293/1818-0442-2024-27-2-64-70
Authors and copyright holders:
- Malkin A. Yu. , National Research Tomsk Polytechnic University (Tomsk, Russia)
- Markov N. G. , National Research Tomsk Polytechnic University (Tomsk, Russia)
- 1. Van Lierop P., Lindquist E., Sathyapala S., Franceschini G. Global forest area disturbance from fire, insect pests, diseases and severe weather events. Forest Ecology and Management, 2015, vol. 352, pp. 78–88.
- 2. Musolin D., Kirichenko N., Karpun N., Mandelshtam M., Selikhovkin A., Zhuravleva E., Aksenenko E., Golub V., Kerchev I., Vasaitis R., Volkovitsh M., Zhuravleva E. Invasive pests of forests and urban trees in Russia: origin pathways, damage, and management. Forests, 2022, vol. 13, p. 521.
- 3. Krivets S.A., Bisirova E.M., Volkova E.S., Debkov N.M., Kerchev I.A., Melnik M.A., Nikiforov A.N., Chernova N.A. Tekhnologiya monitoringa pihtovyh lesov v zone invazii ussurijskogo poligrafa v Sibiri [Stratigraphy and paleogeography of Berriasian-Lower Aptian deposits of Western Siberia in connection with the wedge-like structure of crosssection]. Tomsk, UMIUM Publ., 2018. 74 p. (in Russ.).
- 4. Bystrov S.O., Antonov I.A. First record of the foureyed fir bark beetle Polygraphus proximus Blandford, 1894 (Coleoptera, Curculionidae: Scolytinae) from Irkutsk province Russia. Entomological Review, 2019, vol. 99, pp. 54–55.
- 5. Denisov S.A., Domrachev A.A., Elsukov A.S. Opyt primeneniya kvadrokoptera dlya monitoringa vozobnovleniya lesa [Experience of using quadrocopter for monitoring forest regeneration]. Vestnik of Volga State University of Technology. Series «Forest. Ecology. Nature Managemen», 2016, no. 1 (32), pp. 34–46 (in Russ.).
- 6. Safonova A., Tabik S., Alcaraz-Segura D., Rubtsov A., Maglinets Y., Herrera F. Detection of Fir Trees (Abies sibirica) Damaged by the Bark Beetle in Unmanned Aerial Vehicle Images with Deep Learning. Remote Sensing, 2019, vol. 11, pp. 643–662.
- 7. Xie W., Wang H., Liu W., Zang H. Early-Stage Pine Wilt Disease Detection via Multi-Feature Fusion in UAV Imagery. Forests, 2024, vol. 15, p. 171.
- 8. Kerchev I.A., Maslov K.A., Markov N.G., Tokareva O.S. Semanticheskaya segmentaciya povrezhdennyh derev'ev pihty na snimkah s bespilotnyh letatel'nyh apparatov [Semantic segmentation of damaged fir trees in unmanned aerial vehicle imagery]. Current problems in remote sensing of the Earth from Space, 2021, vol. 18, no. 1, pp. 116–126 (in Russ.).
- 9. Kocon K., Krämer M., Würz H.M. Comparison of CNN-based segmentation models for forest type classification. AGILE: GIScience Series, 2022, vol. 3, p. 42.
- 10. Ronneberger O., Fischer P., Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention, 2015, vol. 9351, pp. 234–241.
- 11. Sivak M.A. Robastnoe obuchenie nejronnyh setej s prostoj arhitekturoj dlya resheniya zadach klassifikacii. Dissertaciya kandidata nauk [Robust learning of neural networks with simple architecture for classification tasks. Dissertation for Candidate of Sciences degree]. Novosibirsk, 2022, 111 p. (in Russ.).
- 12. Markov N.G., Maslov K.A., Kerchev I.A., Tokareva O.S. Modeli U-Net dlya semanticheskoj segmentacii povrezhdennyh derev'ev sosny sibirskoj kedrovoj na snimkah s BPLA [U-Net models for semantic segmentation of damaged Siberian pine cedar trees in UAV imagery]. Current problems in remote sensing of the Earth from Space, 2022, vol. 19, no. 1, pp. 65–77 (in Russ.).
- 13. Clevert D.A., Unterthiner T., Hochreiter S. Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs). Available at: https://arxiv.org/abs/1511.07289, free (Accessed: July 04, 2024).
- 14. Mlotshwa T., van Deventer H., Bosman A.S. Cauchy Loss Function: Robustness Under Gaussian and Cauchy Noise. Available at: https://arxiv.org/abs/2302.07238, free (Accessed: July 04, 2024).
- 15. Lin T., Goyal P., Girshick R., He K., Dollár P. Focal Loss for Dense Object Detection. Available at: https://arxiv.org/abs/1708.02002, free (Accessed: July 04, 2024).
- 16. Bertels J., Eelbode T., Berman M., Vandermeulen D., Maes F., Bisschops R., Blaschko M. Optimizing the Dice Score and Jaccard Index for Medical Image Segmentation: Theory & Practice. Available at: https://arxiv.org/abs/1911.01685, free (Accessed: July 04, 2024).
Last issue
-2026-1_%D0%BA%D0%BE%D0%BF%D0%B8%D1%8F.jpg)
Editorial office address
Executive Secretary of the Editor’s Office
Editor’s Office: 40 Lenina Prospect, Tomsk, 634050, Russia
Phone / Fax: + 7 (3822) 701-582
Viktor N. Maslennikov
Executive Secretary of the Editor’s Office
Editor’s Office: 40 Lenina Prospect, Tomsk, 634050, Russia
Phone / Fax: + 7 (3822) 51-21-21 / 51-43-02
Subscription for updates
