Uneven water distribution in sprinkler irrigation systems, especially due to wind variability and different land slopes, is one of the main problems that reduce irrigation efficiency. In this study, an intelligent approach based
on the integration of GIS (geoinformation systems), machine learning (ML), and deep learning (DL) technologies was developed to solve this problem. In the study, zones were classified into “dry”, “normal”, or “wet” based on zonal parameters such as Slope, Aspect, wind speed and direction, evapotranspiration (ET), and NDVI. An initial classification accuracy of 47.5% was achieved using the Random Forest model. Then, a 1D convolutional neural network (CNN) model was used, to learn learn each zone based on 6 parameters. The CNN model achieved a training accuracy of more than 80% and a validation accuracy of about 75% over 50 epochs. This approach serves as a strong basis for creating an accurate and optimized sprinkler irrigation system under the conditions of climate change and complex terrain in Uzbekistan and Central Asia.
Uneven water distribution in sprinkler irrigation systems, especially due to wind variability and different land slopes, is one of the main problems that reduce irrigation efficiency. In this study, an intelligent approach based
on the integration of GIS (geoinformation systems), machine learning (ML), and deep learning (DL) technologies was developed to solve this problem. In the study, zones were classified into “dry”, “normal”, or “wet” based on zonal parameters such as Slope, Aspect, wind speed and direction, evapotranspiration (ET), and NDVI. An initial classification accuracy of 47.5% was achieved using the Random Forest model. Then, a 1D convolutional neural network (CNN) model was used, to learn learn each zone based on 6 parameters. The CNN model achieved a training accuracy of more than 80% and a validation accuracy of about 75% over 50 epochs. This approach serves as a strong basis for creating an accurate and optimized sprinkler irrigation system under the conditions of climate change and complex terrain in Uzbekistan and Central Asia.
Sprinkler sug‘orish tizimlarida suvning teng taqsimlanmasli-gi sug‘orish samaradorligini pasaytiruvchi asosiy omillardan biri bo‘lib, bunda shamolning o‘zgaruvchanligi va yer nishabligining turlichaligi muhim rol o‘ynay-
di. Ushbu tadqiqotda mazkur muammoni hal qilish maqsadida GIS (geoinformat-sion tizimlar), mashinaviy o‘qitish (ML) va chuqur o‘qitish (DL) texnologiyalari integratsiyasi asosida aqlli yondashuv ishlab chiqildi. Tadqiqotda Slope (nishab-lik), Aspect (yo‘nalish), shamol tezligi va yo‘nalishi, evapotranspiratsiya (ET) va
NDVI kabi parametrlar asosida zonalar “quruq”, “normal” yoki “botqoqlik” ho-latiga ajratildi. Random Forest modeli yordamida dastlabki klassifikatsiyada 47,5 % aniqlikka erishildi. Keyingi bosqichda 1D konvolyutsion neyron tarmoq (CNN) modeli qo‘llanib, bu model har bir zonani 6 ta parametr asosida o‘rgandi.
CNN modeli 50 epoch davomida 80 %dan ortiq o‘quv aniqligi va 75 %ga yaqin tekshiruv aniqligini qayd etdi. Ushbu yondashuv O‘zbekiston va Markaziy Osiyo sharoitida iqlim o‘zgarishi va murakkab relyef sharoitida aniq hamda optimal-lashtirilgan sprinkler sug‘orish tizimini yaratish uchun kuchli asos bo‘lib xizmat qiladi.
Неравномерное распределение воды в системах дождевания, в особенности обусловленное изменчивостью ветровых условий и различиями в уклонах местности, является одной из ключевых проблем, снижающих эффективность орошения. В настоящем исследовании разработан интеллектуальный подход, основанный на интеграции технологий ГИС (геоинформационных систем, GIS), машинного обучения (Machine Learning, ML) и глубокого обучения (Deep Learning, DL), направленный на решение указанной проблемы. В рамках исследования зоны орошения классифицировались как «сухие», «нормальные» или «переувлажнённые» на основе зональных параметров, включающих уклон (Slope), экспозицию склона (Aspect), скорость и направление ветра, значения эвтранспирации (ET) и нормализованный разностный вегетационный индекс (NDVI). На первом этапе с использованием модели Random Forest была достигнута точность классификации 47,5 %. Далее применялась одномерная свёрточная нейронная сеть (1D CNN), обучавшаяся по шести параметрам каждой зоны. Модель CNN обеспечила точность на обучающей выборке более 80 % и точность на валидационной выборке около 75 % при 50 эпохах обучения. Предложенный подход создаёт прочную основу для разработки точной и оптимизированной системы дождевания в условиях изменяющегося климата и сложного рельефа Узбекистана и Центральной Азии.
| № | Муаллифнинг исми | Лавозими | Ташкилот номи |
|---|---|---|---|
| 1 | Arifjanov A.M. | Doctor of Technical Sciences (DSc), Professor | Tashkent Institute of Irrigation and Agricultural Mechanization Engineers” National Research University |
| 2 | Samiev .N. | Doctor of Technical Sciences (DSc), Associate Professor | Tashkent Institute of Irrigation and Agricultural Mechanization Engineers” National Research University |
| 3 | Abduraimova D.A. | Doctor of Philosophy in Technical Sciences (PhD), Associate Professor | Tashkent Institute of Irrigation and Agricultural Mechanization Engineers” National Research University |
| 4 | Jalilova K.A. | Doctoral Student (PhD) | Tashkent Institute of Irrigation and Agricultural Mechanization Engineers” National Research University |
| № | Ҳавола номи |
|---|---|
| 1 | Ababaei, B., & Choo, J. (2022). Smart irrigation: A machine learning and GIS integration approach. Environmental Modelling & Software, 152, 105387. |
| 2 | Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). Crop evapotranspiration – Guidelines for computing crop water requirements (FAO Irrigation and Drainage Paper No. 56). FAO. |
| 3 | Arifjanov, A., Kaletova, T., & Jalilova, K. (2022). Assessment of hydraulic efficiency of irrigation by means of a sprinkler. Science and Innovative Development, 5(6), 26–33. |
| 4 | Arifjanov, A., Samiev, L., & Jalilova, K. (2024). Review of the water distribution modelling in a sprinkler irrigation system in 2000–2020. NL Zhurnal sel’skogo khozyaystva i biotekhnologii [NL Journal of Agriculture and Biotechnology], 1(1), 1–8 |
| 5 | Arifjanov, A., Samiev, L., Abduraimova, D., & Jalilova, K. (2024). Sustainable green development: Water-saving irrigation technologies. E3S Web of Conferences, 574, 05002. https://doi.org/10.1051/ e3sconf/202457405002 |
| 6 | Bai, Y., & Zhou, W. (2021). Deep CNN-based spatial mapping of evapotranspiration from heterogeneous landscapes. ISPRS Journal of Photogrammetry and Remote Sensing, 172, 14–27. |
| 7 | Cao, S., & Lin, H. (2020). Evaluation of sprinkler efficiency using IoT-based sensors and DL algorithms. Sensors, 20(3), 912. |
| 8 | Chen, J., Zhao, M., & Yu, Z. (2019). Application of deep learning in precision irrigation using UAV data. Sensors, 19(17), 3729. |
| 9 | Choudhury, B. J. (1989). Estimating evaporation and transpiration using satellite data. Remote Sensing of Environment, 30(1), 75–91. |
| 10 | Domínguez, A., & Martínez, M. A. (2021). GIS-based modeling for sprinkler design under variable topography. Water, 13(4), 527. |
| 11 | Dong, C., & Wu, Y. (2020). Estimating ET using hybrid ML–DL models with GIS integration. Agricultural Water Management, 239, 106268. |
| 12 | D’Urso, G., & Calera, A. (2022). High-resolution mapping of ET using remote sensing and ML. Agricultural Water Management, 256, 107074. |
| 13 | Esfandiari, M., & Ghasemzadeh, M. (2020). GIS-based multi-criteria decision-making for optimal irrigation system selection. Water, 12(9), 2461. |
| 14 | Evaluation of the Hydraulic Efficiency of the Sprinkler Irrigation System by Christiansen Uniformity Coefficient. (2024). https://doi.org/10.9734/BPI/SRNTA/V1/2050 |
| 15 | Fang, H., & Wang, Y. (2021). Deep spatiotemporal modeling of irrigation distribution. Computers and Electronics in Agriculture, 188, 106348. |
| 16 | Garcia, A., & Tomic, M. (2019). Impact of terrain and meteorological factors on sprinkler performance. Water Resources Management, 33, 2227–2241. |
| 17 | Ghorbani, M. A., & Aalami, M. T. (2019). Wind drift and evaporation loss modeling in sprinkler irrigation systems. Irrigation and Drainage Engineering, 145(2), 04018072. |
| 18 | He, X., & Liang, T. (2019). Convolutional neural network architecture for water stress classification. Remote Sensing, 11(15), 1805 |
| 19 | Hu, J., & Zhang, L. (2022). GIS-based hybrid optimization of irrigation in Central Asia. Journal of Environmental Management, 307, 114569. |
| 20 | . Kang, S., & Zhang, L. (2020). Wind effect correction models for sprinkler application. Journal of Irrigation and Drainage Engineering, 146(9), 04020036. |
| 21 | Karimov, B., & Tursunov, M. (2023). Spatial analysis of irrigation efficiency in Central Asia using RS and GIS. Environmental Earth Sciences, 82, 496 |
| 22 | Li, S., & Shen, Y. (2020). NDVI-ET modeling with random forest and CNN fusion. ISPRS International Journal of Geo-Information, 9(6), 367 |
| 23 | Liang, Y., & Ma, J. (2023). GIS-based DL framework for sprinkler positioning optimization. Remote Sensing, 15(11), 2671 |
| 24 | Liu, Y., Wu, W., & Wang, X. (2021). Evaluation of sprinkler irrigation uniformity under different wind speeds using CFD modeling. Agricultural Water Management, 247, 106728. |
| 25 | Maestre-Valero, J. F., González-Ortega, M. J., & Martínez-Álvarez, V. (2022). Neural network models to optimize sprinkler irrigation under Mediterranean conditions. Agricultural Systems, 195, 103284. |
| 26 | Mohammadi, K., & Anwar, A. H. M. F. (2022). Slope-adjusted water distribution modeling for precision irrigation. Water Resources Research, 58(4), e2021WR031875. |
| 27 | Morales, P., & Camacho, E. (2018). Energy and water optimization in sprinkler systems. Agricultural Systems, 167, 96–106. |
| 28 | Paredes, P., & Teixeira, J. (2022). Remote and in-situ calibration of sprinkler systems using ML. Journal of Hydrology, 608, 127608. |
| 29 | Ramos, T. B., & Albuquerque, A. (2021). Sprinkler uniformity under different wind regimes: Field and model results. Irrigation Science, 39, 371–386. |
| 30 | Rana, G., Katerji, N., & Mastrorilli, M. (2021). Effects of wind and slope on sprinkler distribution patterns. Irrigation Science, 39, 127–138. |
| 31 | Rodríguez-Sinobas, L., & Sánchez-Calvo, R. (2021). Comparative study of surface and sprinkler irrigation uniformity under varying terrain. Journal of Hydrology, 594, 125946. |
| 32 | Salehi, F., & Javan, M. (2021). Random Forest application in irrigation planning: A case study using remote sensing data. Water, 13(8), 1103. |
| 33 | Samiev, L., Jalilova, K., Jalilov, S., Xusanova, J., & Mamatov, D. (2023a). Draft sprinkler irrigation system design development. E3S Web of Conferences, 365, 03024. https://doi.org/10.1051/ e3sconf/202336503024 |
| 34 | Samiev, L., Qurbanov, H., Xusanova, J., Mamatova, D., Djalilov, S., & Jalilova, K. (2023b). Development of technology for reuse of collector trench waters. E3S Web of Conferences, 365, 03031. https://doi.org/10.1051/e3sconf/202336503031 |
| 35 | Shiklomanov, A. I., & Vörösmarty, C. J. (2021). Modeling climate impact on agricultural water use. Nature Reviews Earth & Environment, 2, 678–692. |
| 36 | Solgi, M., & Kamgar-Haghighi, A. A. (2019). Optimizing sprinkler spacing under wind and terrain variability. Irrigation Science, 37, 229–243. |
| 37 | Song, X., & Wang, R. (2021). Multi-layer CNN for spatial irrigation risk detection. Environmental Modelling & Software, 140, 105038. |
| 38 | Tang, H., Zhang, D., & Li, Z. (2023). Integration of GIS and CNN for spatial water allocation modeling. Remote Sensing, 15(4), 852. |
| 39 | Tursunova, N., & Islomov, A. (2023). A GIS–ML–DL integrated framework for sprinkler efficiency optimization in Uzbekistan. Agricultural Water Management. |
| 40 | Wang, L., Zhang, H., & Fang, X. (2021). An assessment of slope impact on sprinkler distribution uniformity. Agricultural Water Management, 243, 106449. |
| 41 | Wang, Y., Chen, J., & Zhao, L. (2020). Modeling wind effects on sprinkler water distribution using machine learning techniques. Water Resources Research, 56(7), e2019WR026715. |
| 42 | Wei, T., & He, Y. (2021). A CNN-LSTM hybrid model for spatial–temporal prediction of irrigation demand. Remote Sensing of Environment, 259, 112418. |
| 43 | Wu, H., & Lin, Y. (2020). Real-time smart irrigation using GIS–DL integration. Sensors, 20(21), 6048. |
| 44 | Xie, Y., & Zhao, X. (2021). Using convolutional autoencoders to predict sprinkler performance zones. Remote Sensing, 13(4), 714 |
| 45 | Yang, D., & Hu, Y. (2023). Slope-aware irrigation modeling using deep learning and GIS. Environmental Earth Sciences, 82(10), 515. |
| 46 | Yazdanpanah, H., & Nazari, R. (2020). A comparative analysis of ML models in predicting water application efficiency. Computers and Electronics in Agriculture, 170, 105266. |
| 47 | Zhang, J., & Liu, W. (2022). Deep residual learning for irrigation pattern prediction. Computers and Electronics in Agriculture, 195, 106877. |
| 48 | Zhang, X., Zhao, H., & Wang, J. (2020). Deep learning model for classifying irrigation zones using Sentinel-2 and DEM. Remote Sensing, 12(5), 745 |
| 49 | Zhang, Y., Zhang, G., & Zhang, X. (2022). Deep learning approach for water distribution prediction in sprinkler systems. Computers and Electronics in Agriculture, 192, 106613 |
| 50 | Zhao, H., Lu, Z., & Lin, H. (2021). CNN-based irrigation scheduling using spatiotemporal NDVI and ET data. Remote Sensing of Environment, 260, 112470. |
| 51 | Zhou, Q., & Zhang, Y. (2022). Comparative study of ML algorithms for zone-wise sprinkler planning. Water, 14(5), 812. |