sibsutisВестник СибГУТИThe Herald of the Siberian State University of Telecommunications and Information Science1998-6920СибГУТИ10.55648/1998-6920-2026-20-2-93-106sibsutis-1079Research ArticleСтатьиАдаптация модели глубокого обучения на основе ResNet для прогнозирования химических свойств почв центральной России по данным спектроскопииAdaptation of a ResNet-based deep learning model for predicting chemical properties of soils in central Russia using spectroscopy datahttps://orcid.org/0009-0006-3920-0183НовоселовИлья ЕвгеньевичNovoselovIlya Evgenievichnovoselov-ie@yandex.ruhttps://orcid.org/0000-0001-8734-2596БаклановаОльга ЕвгеньевнаBaklanovaOlga Evgenievnaoebaklanova@mail.tuФедеральное государственное автономное образовательное учреждение высшего образования «Национальный исследовательский Томский государственный университет»РоссияFederal State Autonomous Educational Institution of Higher Education "National Research Tomsk State University"Russian Federation202622062026202Вестник СибГУТИ93106Copyright © Новоселов И.Е., Бакланова О.Е., 20262026Новоселов И.Е., Бакланова О.Е.Novoselov I.E., Baklanova O.E.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://vestnik.sibsutis.ru/jour/article/view/1079

В данной статье представлена адаптация модели глубокого обучения на основе остаточной нейронной сети ResNet с модулями внимания CBAM для прогнозирования содержания азота, фосфора и калия в почвах центральной России по данным спектроскопии в видимом и ближнем инфракрасном диапазонах. Модель, предобученная на глобальной библиотеке LUCAS, дообучена на локальной выборке из 100 образцов почв Воронежской, Курской, Ростовской областей и Краснодарского края. Применена стратегия частичной заморозки слоев. После дообучения достигнуты коэффициент детерминации 0,91 и относительное отклонение прогноза 3,01 для азота, для фосфора – 0,56 и 2,40 соответственно. Показана возможность использования портативных спектрометров в системах точного земледелия.

The paper presents adaptation of a deep learning model based on residual neural network ResNet with CBAM attention modules for predicting nitrogen, phosphorus and potassium content in soils of central Russia using visible and near-infrared spectroscopy. The model pre-trained on the global LUCAS library was fine-tuned on a local sample of 100 soil samples from Voronezh, Kursk, Rostov regions and Krasnodar Krai. A partial layer freezing strategy was applied. After fine-tuning, the coefficient of determination of 0.91 and the ratio of performance to deviation of 3.01 were achieved for nitrogen, and 0.56 and 2.40 for phosphorus, respectively. The possibility of using portable spectrometers in precision farming systems is shown.

глубокое обучениеостаточные нейронные сетивниманиедообучениепочваспектроскопияdeep learningresidual neural networksattentionfine-tuningsoilspectroscopyReferencesShin, S.K.; Lee, S.J.; Park, J.H. Prediction of Soil Properties Using Vis-NIR Spectroscopy Combined with Machine Learning: A Review. Sensors 2025, 25, 5045. [Электронный ресурс]. DOI: 10.3390/s25165045. URL: https://doi.org/10.3390/s25165045Shin S.K., Lee S.J., Park J.H. Prediction of Soil Properties Using Vis-NIR Spectroscopy Com-bined with Machine Learning: A Review [Prediction of Soil Properties Using Vis-NIR Spectros-copy Combined with Machine Learning: A Review]. Sensors, 2025, vol. 25, article 5045.Сироткин, В. В. Изучение почвенных параметров на основе полевых спектрометрических данных / В. В. Сироткин, С. В. Васюков, Б. М. Усманов // Вестник Удмуртского университета. Серия: Биология. Науки о Земле. – 2020. – Т. 30, № 1. – С. 71–82. [Электронный ресурс]. DOI: 10.35634/2412-9518-2020-30-1-71-82. URL: http://doi.org/10.35634/2412-9518-2020-30-1-71-82Sirotkin V.V., Vasyukov S.V., Usmanov B.M. Izuchenie pochvennykh parametrov na osnove polevykh spektrometricheskikh dannykh [Study of soil parameters based on field spectromet-ric data]. Vestnik Udmurtskogo universiteta. Seriya: Biologiya. Nauki o Zemle, 2020, vol. 30, no. 1, pp. 71–82.Viscarra Rossel, R. A. Using data mining to model and interpret soil diffuse reflectance spectra / R. A. Viscarra Rossel, T. Behrens // Geoderma. – 2010. – Vol. 158, No. 1-2. – P. 46–54. [Электронный ресурс]. DOI: 10.1016/j.geoderma.2009.12.025. URL: https://doi.org/10.1016/j.geoderma.2009.12.025Viscarra Rossel R.A., Behrens T. Using data mining to model and interpret soil diffuse reflec-tance spectra [Using data mining to model and interpret soil diffuse reflectance spectra]. Geoderma, 2010, vol. 158, no. 1-2, pp. 46–54.de Santana, F.B., de Souza, A.M., Poppi, R.J. Visible and near Infrared Spectroscopy Coupled to Random Forest to Quantify Some Soil Quality Parameters. Spectrochim. Acta Part A Mol. Biomol. Spectrosc., 2018, 191, pp. 454–462. [Электронный ресурс]. DOI: 10.1016/j.saa.2017.10.052. URL: https://doi.org/10.1016/j.saa.2017.10.052de Santana F.B., de Souza A.M., Poppi R.J. Visible and near Infrared Spectroscopy Coupled to Random Forest to Quantify Some Soil Quality Parameters [Visible and near Infrared Spec-troscopy Coupled to Random Forest to Quantify Some Soil Quality Parameters]. Spectrochim. Acta Part A Mol. Biomol. Spectrosc., 2018, vol. 191, pp. 454–462.Chen, S. Rapid determination of soil classes in soil profiles using vis–NIR spectroscopy and multiple objectives mixed support vector classification / S. Chen, S. Li, W. Ma [et al.] // European Journal of Soil Science. – 2019. – Vol. 70, No. 1. – P. 42–53. [Электронный ресурс]. DOI: 10.1111/ejss.12715. URL: https://doi.org/10.1111/ejss.12715Chen S., Li S., Ma W. et al. Rapid determination of soil classes in soil profiles using vis–NIR spectroscopy and multiple objectives mixed support vector classification [Rapid determination of soil classes in soil profiles using vis–NIR spectroscopy and multiple objectives mixed sup-port vector classification]. European Journal of Soil Science, 2019, vol. 70, no. 1, pp. 42–53.Tóth, G. LUCAS Topsoil Survey. Methodology, data and results / G. Tóth, A. Jones, L. Montanarella (eds.). – Luxembourg: Publications Office of the European Union, 2013. – 141 p. – (EUR 26102). [Электронный ресурс]. DOI: 10.2788/97922. https://doi.org/10.2788/97922Tóth G., Jones A., Montanarella L. (eds.). LUCAS Topsoil Survey. Methodology, data and re-sults [LUCAS Topsoil Survey. Methodology, data and results]. Luxembourg, Publications Of-fice of the European Union, 2013. 141 p. (EUR 26102).Ng, W. Kattenborn, T., Leitloff, J., Schiefer, F., Hinz, S. Review on Convolutional Neural Networks (CNN) in Vegetation Remote Sensing. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 173, pp. 24–49. [Электронный ресурс]. DOI: 10.1016/j.isprsjprs.2020.12.010. URL: https://doi.org/10.1016/j.isprsjprs.2020.12.010Kattenborn T., Leitloff J., Schiefer F., Hinz S. Review on Convolutional Neural Networks (CNN) in Vegetation Remote Sensing [Review on Convolutional Neural Networks (CNN) in Vegetation Remote Sensing]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, vol. 173, pp. 24–49.He, K. Deep Residual Learning for Image Recognition / K. He, X. Zhang, S. Ren, J. Sun // 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). – Las Vegas, NV, USA, 2016. – P. 770–778. [Электронный ресурс]. DOI: 10.48550/arXiv.1512.03385. URL: https://doi.org/10.48550/arXiv.1512.03385He K., Zhang X., Ren S., Sun J. Deep Residual Learning for Image Recognition [Deep Residual Learning for Image Recognition]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 2016, pp. 770–778.Wartini Ng. Convolutional neural network for simultaneous prediction of several soil properties using visible/near-infrared, mid-infrared, and their combined spectra/ Wartini Ng, Budiman Minasny // Geoderma. – 2019. – Vol. 352. – P. 251–267. [Электронный ресурс]. DOI: 1016/j.geoderma.2019.06.016. URL: https://doi.org/10.1016/j.geoderma.2019.06.016Ng W., Minasny B. Convolutional neural network for simultaneous prediction of several soil properties using visible/near-infrared, mid-infrared, and their combined spectra [Convolutional neural network for simultaneous prediction of several soil properties using visible/near-infrared, mid-infrared, and their combined spectra]. Geoderma, 2019, vol. 352, pp. 251–267.Woo, S Woo, S., Park, J., Lee, J.-Y., Kweon, I.S. CBAM: Convolutional Block Attention Module. ArXiv, 2018, art. 1807.06521. [Электронный ресурс]. DOI: 10.48550/arXiv.1807.06521. URL: https://doi.org/10.48550/arXiv.1807.06521Woo S., Park J., Lee J.-Y., Kweon I.S. CBAM: Convolutional Block Attention Module [CBAM: Convolutional Block Attention Module]. ArXiv, 2018, art. 1807.06521.Wang, X. Simultaneous estimation of multiple soil properties from vis-NIR spectra using a multi-gate mixture-of-experts with data augmentation / X. Wang, Mei-Wei Zhang, Ya-Nan Zhou [et al.] // Geoderma. – 2024. – Vol. 445. – Article no. 116888. [Электронный ресурс]. DOI: 10.1016/j.geoderma.2024.117127. URL: https://doi.org/10.1016/j.geoderma.2024.117127Wang X., Zhang M.-W., Zhou Y.-N. et al. Simultaneous estimation of multiple soil properties from vis-NIR spectra using a multi-gate mixture-of-experts with data augmentation [Simultane-ous estimation of multiple soil properties from vis-NIR spectra using a multi-gate mixture-of-experts with data augmentation]. Geoderma, 2024, vol. 445, article 116888.Zhong, L. Soil properties: Their prediction and feature extraction from the LUCAS spectral library using deep convolutional neural networks / L. Zhong, X. Guo, Z. Xu, M. Ding // Geoderma. – 2021. – Vol. 402. – Article no. 115366. [Электронный ресурс]. DOI: 10.1016/j.geoderma.2021.115366. URL: https://doi.org/10.1016/j.geoderma.2021.115366Zhong L., Guo X., Xu Z., Ding M. Soil properties: Their prediction and feature extraction from the LUCAS spectral library using deep convolutional neural networks [Soil properties: Their prediction and feature extraction from the LUCAS spectral library using deep convolutional neural networks]. Geoderma, 2021, vol. 402, article 115366.Новоселов, И. Е. Программная реализация модели глубокого обучения для прогнозирования свойств почвы на основе данных спектроскопии / И. Е. Новоселов, О. Е. Бакланова // Вестник Российского нового университета. Серия: Сложные системы: модели, анализ, управление. – 2025. – № 3. – С. 125–139. [Электронный ресурс]. DOI: 10.18137/RNUV9187.25.03.P.125. URL: https://vestnik-rosnou.ru/сложные-системы-модели-анализ-и-управление-complex-systems-models-analysis-management/2025/3/125?language=enNovoselov I.E., Baklanova O.E. Programmnaya realizatsiya modeli glubokogo obucheniya dlya prognozirovaniya svoystv pochvy na osnove dannykh spektroskopii [Software implemen-tation of a deep learning model for predicting soil properties based on spectroscopy da-ta]. Vestnik Rossiyskogo novogo universiteta. Seriya: Slozhnye sistemy: modeli, analiz, uprav-lenie, 2025, no. 3, pp. 125–139.Буряк, Ж. А. Оценка спектрально-отражательных свойств эродированных агропочв Республики Татарстан / Ж. А. Буряк, А. М. Гафуров // Региональные геосистемы. 49, 3 (сен. 2025), 517-532. [Электронный ресурс]. DOI: 10.52575/2712-7443-2025-49-3-517-532. URL: http://doi.org/10.52575/2712-7443-2025-49-3-517-532Buryak Zh.A., Gafurov A.M. Otsenka spektral'no-otrazhatel'nykh svoystv erodirovannykh ag-ropochv Respubliki Tatarstan [Assessment of spectral-reflective properties of eroded agricul-tural soils of the Republic of Tatarstan]. Regional'nye geosistemy, 2025, vol. 49, no. 3, pp. 517-532.Горохова, И. Н. Распознавание каменистых, песчаных и карбонатных с поверхности почв на юге приволжской возвышенности (волгоградская область) по космическим снимкам / И. Н. Горохова, Н. Б. Хитров // Почвоведение. 2023, № 11, с. 1340–1356. [Электронный ресурс]. DOI: 10.31857/S0032180X23600609. URL: https://www.elibrary.ru/download/elibrary_54403217_55348489.pdfGorokhova I.N., Khitrov N.B. Raspoznavanie kamenistykh, peschanykh i karbonatnykh s poverkhnosti pochv na yuge Privolzhskoy vozvyshennosti (Volgogradskaya oblast') po kos-micheskim snimkam [Recognition of stony, sandy, and calcareous surface soils in the south of the Volga Upland (Volgograd region) from satellite images]. Pochvovedenie, 2023, no. 11, pp. 1340–1356.

The authors declare that there are no conflicts of interest present.