| Issue № 4 |
Methods of ecological investigations |
pdf-version |
Using the Keras library to decrypt animal habitats using deep learning methods
| Marfitsyna Natalya Alexandrovna | Petrozavodsk State University, 33 Lenin Ave., Petrozavodsk, Republic of Karelia, 185910, Russia, marfitsyna.nata@mail.ru |
| Korosov Andrey Victorovich | DSc, professor, Petrozavodsk State University, 33 Lenin Ave., Petrozavodsk, Republic of Karelia, 185910, Russia, korosov@psu.karelia.ru |
|
Keywords: habitat RSD GIS neural network R Keras |
Summary: The paper considers the use of deep learning algorithms from the Keras library to solve the problem of classifying forest clearings of different ages using remote sensing in the R environment. A rather complicated procedure for installing Keras libraries on a computer is considered in detail. The stages of neural simulation and their variations using the R neuralnet package and the Keras environment are described. Satellite images were decoded in the vicinity of Gomselga village (Karelia) using field survey data. The typical decryption algorithm (classification with learning) was supplemented by a joint multidimensional analysis of the brightness characteristics of the image and field geobotanical descriptions. As a result, 4 sets of reference signatures were formed, corresponding to a particular state of regenerating clearings. The neural network (multilayer perceptron) was configured to recognize these types of plantings, and then performed the classification of the remaining pixels of the image for the entire studied area. Based on the analysis of geobotanical descriptions and satellite data, a grid map was created highlighting four main types of habitats: fresh cuttings, regenerating cuttings, young trees, and deciduous forest. Data processing using Keras algorithms significantly speeds up analysis, and makes it possible to increase the number of layers and neurons and detail the grid. In particular, unlike the algorithms of reference decoding, the proposed approach made it possible to identify the heterogeneity of vegetation within the same-age clearings. The results of the work are used to identify heterogeneous animal habitats and the influence of environmental factors on their spatial distribution and abundance. © Petrozavodsk State University |
Received on: 16 October 2025 Published on: 27 December 2025 | |
References
Baluta V. I. Osipov V. P. Rykov Yu. G. Chetverushkin B. N. On the notion of influence in the concept of cognitive modeling when using a ReLU-type activation function, Informacionnye tehnologii i vychislitel'nye sistemy. 2023. No. 4. P. 59–71. URL: https://www.elibrary.ru/download/elibrary_56573800_56563851.pdf (data obrascheniya: 08.09.2025).
Bugmyrin S. N. Korosov A. V. Ieshko E. P. Anikanova V. S. Bespyatova L. A. Matrosova Yu. M. Telegin I. V. Experience in studying the spatial distribution of small mammalian parasites, Severnaya Evropa v XXI veke: Priroda, kul'tura, ekonomika: Materialy mezhdunar. konf., posvyasch. 60-letiyu KarNC RAN (24–27 oktyabrya 2006). Petrozavodsk, 2006. P. 55–58.
Danilova I. V. Korec M. A. Ryzhkova V. A. Mapping of age stages of forest vegetation based on the analysis of multi-seasonal satellite images by Landsat, Issledovanie zemli iz kosmosa. 2017. No. 4. P. 12–24. URL: https://elibrary.ru/download/elibrary_29992173_79574715.pdf (data obrascheniya: 10.03.2025).
Eleshkevich A. D. Eremenko M. S. Saybel' E. G. Hristolyubov I. A. Chernov A. G. Application of automated decryption methods in the tasks of recognizing anthropogenic impact zones in oil and gas fields, Ekspoziciya Neft' Gaz. 2023. No. 7. P. 127–131. DOI: 10.24412/2076-6785-2023-7-127-131
GeeksforGeeks. How to Install Keras in Windows. 2021. URL: https://www.geeksforgeeks.org/how-to-install-keras-in-windows/?ysclid=m8oygmq0nq900158163 (data obrascheniya: 15.01.2025).
Geoinformatics: A textbook for university students, E. G. Kapralov, A. V. Koshkarev, V. P. Tikunov i dr.; Pod red. V. P. Tikunova. M.: Izd. centr «Akademiya», 2005. 480 p.
Guseva T. L. Korosov A. V. Bespyatova L. A. Anikanova V. S. Long-term dynamics of biotopic distribution of the common shrew (Sorex araneus, Linnaeus 1758) in mosaic landscapes of Karelia, Uchenye zapiski Petrozavodskogo gosudarstvennogo universiteta. 2014. T. 2, No. 8. P. 13–20. URL: https://cyberleninka.ru/article/n/mnogoletnyaya-dinamika-biotopicheskogo-razmescheniya-obyknovennoy-burozubki-sorex-araneus-linnaeus-1758-v-mozaichnyh-landshaftah/viewer (data obrascheniya: 10.03.2025).
Ieshko E. P. Korosov A. V. Nikonorova I. A. Bugmyrin S. V. The relationship between the species richness of helminth communities and the abundance of the host (using the example of the common shrew Soxrex araneus), Parazitologiya. 2020. T. 54, No. 1. P. 3–12.
Il'yuchik M. A. Cay S. S. Formation of a database of reference and calibration sites for thematic decoding of space survey materials, Trudy BGTU. Seriya 1. Lesnoe hozyaystvo. 2010. Vyp. 18. P. 44–47.
Korosov A. V. Marficyna N. A. Decryption of animal habitats using deep learning methods of Keras Library, InterKarta. InterGIS: Materialy mezhdunar. konf. Perm', 2025. (V pechati)
Korosov A. V. Matrosova Yu. M. Bugmyrin S. V. Anikanova B. C. Bespyatova L. A. The experience of reconstruction of the territorial location of the red vole in the mosaic landscape of South Karelia, Biogeografiya Karelii (flora i fauna taezhnyh ekosistem). Vyp. 4. Petrozavodsk, 2003. P. 204–212.
Korosov A. V. Neural networks for ecology: an introduction, Principy ekologii. 2023. No. 3. P. 76–96. DOI: 10.15393/j1.art.2023.14002 (data obrascheniya: 26.02.2025).
Kosulin V. V. Application of artificial intelligence systems in environmental monitoring of enterprise emissions, Nauchno-tehnicheskiy vestnik Povolzh'ya. 2023. No. 10. P. 132–135. URL: elibrary_54795395_27315499.pdf (data obrascheniya: 26.02.2025).
Kuz'menko E. I. Frolov A. A. Silaev A. V. Mapping of forest landscapes in the north-west of Western Siberia using GIS, Geografiya i prirodnye resursy. 2015. No. 4. P. 151–161.
Lavrinenko I. A. Modern approaches to the preparation of maps of Arctic vegetation, Geobotanicheskoe kartografirovanie. 2023. P. 14–27. URL: https://cyberleninka.ru/article/n/sovremennye-podhody-k-podgotovke-kart-rastitelnosti-arktiki (data obrascheniya: 12.12.2025).
Lavrinenko I. A. Typology of territorial units of vegetation for the purposes of large-scale mapping (using the example of the Island of Kolguyev ), Geobotanicheskoe kartografirovanie. 2015. P. 95–119. URL: https://cyberleninka.ru/article/n/tipologiya-territorialnyh-edinits-rastitelnosti-dlya-tseley-krupnomasshtabnogo-kartografirovaniya-na-primere-ostrova-kolguev (data obrascheniya: 12.12.2025).
PythonRu. 2021. URL: https://pythonru.com/biblioteki/kak-ustanovit-keras-na-linux-i-windows-keras?ysclid=m8co0qtcos915911868 (data obrascheniya: 15.01.2025).
Raevskiy B. V. Tarasenko V. V. Petrov N. V. Assessment of the current state of the plant communities of the Kostomukshsky Nature Reserve using satellite images from the Landsat system, Sovremennye problemy distancionnogo zondirovaniya Zemli iz kosmosa. 2022. T. 19, No. 3. P. 47–61. URL: http://jr.rse.cosmos.ru/article.aspx?id=2528 (data obrascheniya: 10.11.2025).
Semakina A. V. Shihov A. N. Klimina E. A. Mapping the vulnerability of forests to wind impacts based on satellite data (using the example of the Perm Region), Sovremennye problemy distancionnogo zondirovaniya Zemli iz kosmosa. 2025. T. 22, No. 4. P. 236–252.
Shitikov V. K. Mastickiy S. E. Classification, regression, and other Data Mining algorithms using R. 2017. 351 p. URL: https://github.com/ranalytics/data-mining (data obrascheniya: 26.01.2025).
Sholle F. Deep learning with R and Keras. M.: DMK Press, 2022. 646 p. URL: https://coollib.net/b/627871-fransua-sholle-glubokoe-obuchenie-s-r-i-keras (data obrascheniya: 26.02.2025).
Sochilova E. N. Ershov D. V. Analysis of the possibility of determining tree stocks based on Landsat ETM satellite data+, Sovremennye problemy distancionnogo zondirovaniya Zemli iz kosmosa. 2012. T. 9, No. 3. P. 277–282. URL: http://jr.rse.cosmos.ru/article.aspx?id=1076 (data obrascheniya: 26.02.2025).
The R Project for Statistical Computing. 2023. URL: https://www.r-project.org/ (data obrascheniya: 26.01.2025).
© 2011 - 2026
Petrozavodsk State University
technical support - RCNIT
The editor-in-chief - Andrey Korosov
ecopri@petrsu.ru