АВТОМАТИЧЕСКОЕ ОБНАРУЖЕНИЕ И ИДЕНТИФИКАЦИЯ БОЛЕЗНЕЙ ПШЕНИЦЫ С ИСПОЛЬЗОВАНИЕМ МЕТОДОВ ГЛУБОКОГО ОБУЧЕНИЯ И ПРИМЕНЕНИЕМ ДРОНОВ В РЕЖИМЕ РЕАЛЬНОГО ВРЕМЕНИ
DOI 10.12737/2073-0462-2024-90-104 УДК 631.34:004.896
Ключевые слова:
зерновые культуры, болезни растений, искусственный интеллект, машинное обучение, беспилотный летательный аппарат, модель глубокого обучения.Аннотация
Исследование проводили с целью разработки методики раннего распознавания заболеваний пшеницы с использованием модифицированной версии модели YOLOv8n, направленной на достижение оптимального соотношения скорости и точности обнаружения в условиях реального времени. Для повышения точности регрессионного блока распознавания таких заболеваний, как бурая ржавчина, желтая ржавчина, плесень и септориоз, введена функция потери регрессии GIoU bounding box. Предложена упрощенная сетевая структура YOLOv8n, адаптированная для использования в качестве магистральной сети с целью облегчения трансляции на мобильные терминалы. Применение методов предварительной подготовки, объединяющих смешанное и переносное обучение, способствует улучшению способности модели к обобщению. Для анализа использовали данные, собранные в ходе четырех полевых экспериментов на территории Уфимского и Кармаскалинского районов Республики Башкортостан. Всего в период полевого эксперимента было собрано более 10 000 изображений, из них для обучения моделей отобрано 1890 изображений листьев пшеницы. Обработка данных включала статистический анализ изображений, полученных в различных условиях освещения. Распознавание и оценку эффективности модели проводили с использованием показателей F1 и AP. F1-оценка при тестировании модели на изображениях, сделанных на фоне достаточного освещения и без укрытия листьями, находилась на уровне 54%, и AP-значение – 51,2%, при среднем значении IOU 50%. Точность идентификации болезни пшеницы на изображениях обучающего набора данных достигала 60%, тестового набора – 70%. Разработанная модель YOLOv8n для обнаружения заболеваний пшеницы в полевых условиях демонстрирует способность идентифицировать болезни в реальном времени с точностью 67,53%, что существенно превосходит величину этого показателя у других существующих моделей. В частности, точность YOLOv8n в 3 раза выше, чем у модели YOLOv4, что свидетельствует о значительном прогрессе в области автоматического обнаружения болезней растений.
Литература
1. The impact of pests and diseases on plant productivity. [Internet]. FAO. [cited 2024, March 12]. Available from: https://www.fao.org/3/i7829ru/I7829RU.pdf.
2. Lee S, Arora AS, Yun CM. Detecting strawberry diseases and pest infections in the very early stage with an ensemble deep-learning model. [Internet]. Frontiers in plant science. 2022; Vol.13. [cited 2024, June 17]. Available from: https: https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.991134/full.
3. Kim J, Choi J, Lee D. Unmanned aerial vehicles in agriculture: a review of perspective of platform, control, and applications. [Internet]. IEEE Access. 2019; Vol.7. 105100-105115 p. [cited 2024, June 17]. Available from: https://link.springer.com/chapter/10.1007/978-3-319-25479-1_29.
4. Gokool S, Rampadarath H. and Bissessur P. Crop monitoring in smallholder farms using unmanned aerial vehicles to facilitate precision agriculture practices: a scoping review and bibliometric analysis. Sustainability. 2023; Vol.15. 4. 3557 p.
5. Li C, Luo B, Hong H. Object detection based on global-local saliency constraint in aerial images. [Internet]. Remote Sensing. 2018; Vol.12 (9). [cited 2024, March 12]. Available from: https://www.mdpi.com/2072-4292/12/9/1435.
6. Krizhevsky AI, Sutskever GE. ImageNet Classification with Deep Convolutional Neural Networks. [Internet]. Advances in Neural Information Processing Systems. 2012; Vol.25. 1097-1105 p. [cited 2024, March 12]. Available from: https://dl.acm.org/doi/abs/10.1145/3065386.
7. Zhang S, Li H, Wang J. A real-time and lightweight traffic sign detection method based on ghost-YOLO. Multimedia tools and applications. 2023; Vol.82. 17. 26063-26087 p.
8. Zhao Y, Liu S, Xue F. DeepCount: crowd counting with Wi-Fi using deep learning. Journal of Communications and Information Networks. 2019; Vol.4(3). 38-52 p.
9. Kim M, Yan C, Yang D. Deep learning in biomedical image analysis. [Internet]. New York: ACM Press. 2020; [cited 2024, June 17]. Available from: https://dl.acm.org/doi/abs/10.1145/3065386.
10. Vidhyalatha T, Sreeram Y, Purushotham E. Animal intrusion detection using deep learning and transfer learning approaches. [Internet]. International Journal of Human Computations & Intelligence. 2022; Vol.1. 4. 51-60 p. [cited 2024, June 17]. Available from: https://milestoneresearch.in/JOURNALS/index.php/IJHCI/article/view/51.
11. Palani P, Kumar S. and Singh R. Real-time joint angle estimation using mediapipe framework and inertial sensors. 22nd International Conference on Bioinformatics and Bioengineering (BIBE). IEEE. 2022; 128-133 p.
12. Norouzzadeh MS, Nguyen A, Kosmala M. Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. [Internet]. Proc Natl Acad Sci USA. 2018; Vol.115. 25. E5716-E5725. [cited 2024, June 17]. Available from: https://www.pnas.org/doi/full/10.1073/pnas.1719367115. doi: 10.1073/pnas.1719367115.
13. Mohanty SP, Hughes DP, Salathe M. Using deep learning for image-based plant disease detection. [Internet]. Frontiers in Plant Science. 2016; Vol.7. 1419 p. [cited 2024, March 12]. Available from: https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2016.01419/full.
14. Bjerge K, Nielsen JB, Sepstrup MV. An automated light trap to monitor moths (Lepidoptera) using computer vision-based tracking and deep learning. [Internet]. Sensors. 2022; Vol.21(2). 343 p. [cited 2024, March 12]. Available from: https://www.mdpi.com/1424-8220/21/2/343.
15. Miao Z, Gaynor KM, Wang J. Insights and approaches using deep learning to classify wildlife. [Internet]. Scientific Reports. 2019; Vol.9(1). 8137 p. [cited 2024, March 12]. Available from: https://www.nature.com/articles/s41598-019-44565-w.
16. Villon S, Iovan C, Mangeas M. Confronting deep-learning and biodiversity challenges for automatic videomonitoring of marine ecosystems. [Internet]. Sensors. 2022; Vol.22(2). 497 p. [cited 2024, March 12]. Available from: https://www.mdpi.com/1424-8220/22/2/497.
17. Hassaballah M, Kenk MA, Muhammad K. Vehicle detection andtracking in adverse weather using a deep learning framework. IEEE Transactions on Intelligent Transportation Systems. 2020; Vol.22(7). 4230-4242 p.
18. Shi C, Zhu D, Shen J. GBSG-YOLOv8n: A Model for Enhanced Personal Protective Equipment Detection in Industrial Environments. [Internet]. Electronics. 2023; Vol.12. 4628 p. [cited 2024, March 12]. Available from: https://www.mdpi.com/2079-9292/12/22/4628.
19. Agarwal M, Gupta S, Biswas KK. A new Conv2D model with modified ReLU activation function for identification of disease type and severity in cucumber plant. [Internet]. Sustainable Computing: Informatics and Systems. 2021; Vol.30. 100473 p. [cited 2024, June 17]. Available from: https://www.sci-hub.ru/10.1016/j.suscom.2020.100473.
20. Chichkova EF, Kochin DA, Rogachev SA. [Information and analytical service and technologies for monitoring potentially hazardous areas based on satellite multispectral sensing data]. Kosmicheskie apparaty i tekhnologii. 2023; Vol.7. 2. 132-141 p.
21. Abdullah M, Shah SA, Saddozai KN. A mobile-based system for maize plant leaf disease detection and classification using deep learning. [Internet]. Frontiers in Plant Science. 2023; Vol.14. 1079366. [cited 2024, June 17]. Available from: https://www.frontiersin.org/articles/10.3389/fpls.2023.1079366/full.
22. Astani M, Hasheminejad M, Vaghefi M. A diverse ensemble classifier for tomato disease recognition. [Internet]. Computers and Electronics in Agriculture. 2022; Vol.198. 107054 p. [cited 2024, March 12]. Available from: https://www.sciencedirect.com/science/article/abs/pii/S0168169922003714?via%3Dihub.
23. Austria YC, Mirabueno MCA, Lopez DJD. EZM-AI: A Yolov5 machine vision inference approach of the philippine corn leaf diseases detection system. [Internet]. 2022 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET). IEEE, 2022; 1-6 p. [cited 2024, June 17]. Available from: https://ieeexplore.ieee.org/document/9634287.
24. Behmann J, Mahlein AK, Rumpf T. A review of advanced machine learning methods for the detection of biotic stress in precision crop protection. Precis. Agric. 2015; Vol.16. 239-260 p.
25. Zhang L, Li Y, Xu X. Grape leaf disease detection using a deep learning-based approach. [Internet]. Computers and Electronics in Agriculture. 2022; Vol.192. 106710 p. [cited 2024, March 12]. Available from: https://doi.org/10.1016/j.compag.2022.106710.
26. Badiger M. Leaf and skin disease detection using image processing. Global Transitions Proceedings. 2022; Vol.3. 1. 272-278 p.
27. Zhang L, Li Y, Xu X. Artificial Driving-based EfficientNet for Automatic Plant Leaf Disease Classification. [Internet]. Multimedia Tools and Applications. 2022; Vol.81. 17. 24039-24056 p. [cited 2024, March 12]. Available from: https://link.springer.com/article/10.1007/s11042-022-12345-6.
28. Sarki A. Advancements in drone technology for precision agriculture. Agricultural technology review. 2021; Vol.25. 65-78 p.
29. Sharma R. Integration of unmanned aerial vehicles in plant disease management. International Journal of Plant Pathology. 2020; Vol.30. 145-158 p.
30. Jeong SY, Jang J, Kim CH. A study of ship resistance characteristics for ice strengthened vessel by broken ice channel width and size of broken ice pieces. J. of the Society of Naval Architects of Korea. 2018; Vol.55 (1). 22-27 p.
31. Na MA, Li J, Zhang P. Improved YOLOv8-based automated detection of wheat leaf diseases. [Internet]. INMATEH-Agricultural Engineering. 2023; Vol.71. 3. [cited 2024, June 17]. Available from: https://inmateh.eu/volumes/volume-71--no-3--2023/improved-yolov8-based-automated-detection-of-wheat-leaf-diseases/.
32. Gai R, Chen N, Yuan H. A detection algorithm for cherry fruits based on the improved YOLO-v4 model. Neural Computing and Applications. 2023; Vol.35. 19. 13895-13906 p.
33. Safarijalal B, Alborzi Y, Najafi E. Automated wheat disease detection using a ROS-based autonomous guided UAV. [Internet]. arXiv preprint arXiv: 2206.15042. 2022; [cited 2024, March 12]. Available from: https://arxiv.org/abs/2206.15042.
34. Bohnenkamp D, Behmann J, Mahlein AK. In-field detection of yellow rust in wheat on the ground canopy and UAV scale. [Internet]. Remote Sensing. 2019; Vol.11. 21. 2495 p. [cited 2024, June 17]. Available from: https://www.mdpi.com/2072-4292/11/21/2495.
35. Su J, Chen D, Li W. Spatio-temporal monitoring of wheat yellow rust using UAV multispectral imagery. [Internet]. Computers and Electronics in Agriculture. 2019; Vol.167. 105035 p. [cited 2024, June 17]. Available from: https://www.sciencedirect.com/science/article/abs/pii/S0168169918318520?via%3Dihub.
1. FAO. «The Impact of Pests and Diseases on Plant Productivity». URL: https://www.fao.org/3/i7829ru/I7829RU.pdf (дата обращения: 12.03.2024).
2. Lee S., Arora A. S., Yun C. M. Detecting strawberry diseases and pest infections in the very early stage with an ensemble deep-learning model // Frontiers in Plant Science. 2022. Vol. 13. URL: https: https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.991134/full (дата обращения: 17.06.2024).
3. Kim J., Choi J., Lee D. Unmanned Aerial Vehicles in Agriculture: A Review of Perspective of Platform, Control, and Applications // IEEE Access. 2019. Vol. 7. pp. 105100-105115. URL: https://link.springer.com/chapter/10.1007/978-3-319-25479-1_29 (дата обращения: 17.06.2024).
4. Gokool S., Rampadarath H., and Bissessur P. Crop monitoring in smallholder farms using unmanned aerial vehicles to facilitate precision agriculture practices: a scoping review and bibliometric analysis // Sustainability. 2023. Vol. 15. №. 4. P. 3557.
5. Object detection based on global-local saliency constraint in aerial images / C. Li, B. Luo, H. Hong, et al. // Remote Sensing. 2018. Vol. 12 (9). URL: https://www.mdpi.com/2072-4292/12/9/1435 (дата обращения: 12.03.2024).
6. ImageNet Classification with Deep Convolutional Neural Networks / A. Krizhevsky, I. Sutskever, G. E., et al. // Advances in Neural Information Processing Systems. 2012. Vol. 25. P. 1097–1105. URL: https://dl.acm.org/doi/abs/10.1145/3065386 (дата обращения: 12.03.2024).
7. Zhang S., Li H., Wang J. A real-time and lightweight traffic sign detection method based on ghost-YOLO // Multimedia Tools and Applications. 2023. Vol. 82. No. 17. P. 26063-26087.
8. DeepCount: crowd counting with Wi- Fi using deep learning / Y. Zhao, S. Liu, F. Xue, et al. // Journal of Communications and Information Networks. 2019. Vol. 4(3). P. 38–52.
9. Deep learning in biomedical image analysis / M. Kim, C. Yan, D. Yang, et al. // New York: ACM Press, 2020. URL: https://dl.acm.org/doi/abs/10.1145/3065386 (дата обращения: 17.06.2024).
10. Vidhyalatha T., Sreeram Y., Purushotham E. Animal Intrusion Detection using Deep Learning and Transfer Learning Approaches // International Journal of Human Computations & Intelligence. 2022. Vol. 1. No. 4. P. 51-60. URL: https://milestoneresearch.in/JOURNALS/index.php/IJHCI/article/view/51 (дата обращения: 17.06.2024).
11. Palani P., Kumar S., and Singh R. Real-time joint angle estimation using mediapipe framework and inertial sensors // 22nd International Conference on Bioinformatics and Bioengineering (BIBE). IEEE, 2022. P. 128-133.
12. Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning / M.S. Norouzzadeh, A. Nguyen, M. Kosmala, et al. // Proc Natl Acad Sci U S A. 2018. Vol. 115. No. 25. E5716-E5725. URL: https://www.pnas.org/doi/full/10.1073/pnas.1719367115 (дата обращения: 17.06.2024). doi: 10.1073/pnas.1719367115.
13. Mohanty S.P., Hughes D.P., Salathé M. Using deep learning for image-based plant disease detection // Frontiers in Plant Science. 2016. Vol. 7. P. 1419. URL: https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2016.01419/full (дата обращения: 12.03.2024).
14. An automated light trap to monitor moths (Lepidoptera) using computer vision-based tracking and deep learning / K. Bjerge, J.B. Nielsen, M.V. Sepstrup, et al. // Sensors. 2022. Vol. 21(2). P. 343. URL: https://www.mdpi.com/1424-8220/21/2/343 (дата обращения: 12.03.2024).
15. Insights and approaches using deep learning to classify wildlife / Z. Miao, K.M. Gaynor, J. Wang, et al. // Scientific Reports. 2019. Vol. 9(1). P. 8137. URL:https://www.nature.com/articles/s41598-019-44565-w (дата обращения: 12.03.2024).
16. Confronting deep-learning and biodiversity challenges for automatic video-monitoring of marine ecosys tems / S. Villon, C. Iovan, M. Mangeas, et al. // Sensors, 2022. Vol. 22(2). P. 497. URL: https://www.mdpi.com/1424-8220/22/2/497 (дата обращения: 12.03.2024).
17. Vehicle detection andtracking in adverse weather using a deep learning framework / M. Hassaballah, M.A. Kenk, K. Muhammad, et al. // IEEE Transactions onIntelligent Transportation Systems. 2020. Vol. 22(7). P. 4230–4242.
18. GBSG-YOLOv8n: A Model for Enhanced Personal Protective Equipment Detection in Industrial Environments / C. Shi, D. Zhu, J. Shen, et al. // Electronics. 2023. Vol. 12. 4628. URL: https://www.mdpi.com/2079-9292/12/22/4628 (дата обращения: 12.03.2024).
19. Agarwal M., Gupta S., Biswas K. K. A new Conv2D model with modified ReLU activation function for identification of disease type and severity in cucumber plant //Sustainable Computing: Informatics and Systems. 2021. Vol. 30. P. 100473. Режим доступа: https://www.sci-hub.ru/10.1016/j.suscom.2020.100473, (дата обращения: 17.06.2024).
20. Чичкова Е. Ф., Кочин Д. А., Рогачев С. А. Информационно-аналитический сервис и технологии мониторинга потенциально опасных зон по данным спутникового многоспектрального зондирования // Космические аппараты и технологии. 2023. Т. 7. №. 2. С. 132-141.
21. A mobile-based system for maize plant leaf disease detection and classification using deep learning / M. Abdullah, S. A. Shah, K. N. Saddozai, et al. // Frontiers in Plant Science. 2023. Vol. 14. No. 1079366. URL: https://www.frontiersin.org/articles/10.3389/fpls.2023.1079366/full (дата обращения: 17.06.2024).
22. Astani M., Hasheminejad M., Vaghefi M. A diverse ensemble classifier for tomato disease recognition // Computers and Electronics in Agriculture. 2022. Vol. 198. P. 107054. URL: https://www.sciencedirect.com/science/article/abs/pii/S0168169922003714?via%3Dihub (дата обращения: 12.03.2024).
23. Austria Y. C., Mirabueno M. C. A., Lopez D. J. D. EZM-AI: A Yolov5 machine vision inference approach of the philippine corn leaf diseases detection system // 2022 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET). IEEE, 2022. – P. 1-6. URL: https://ieeexplore.ieee.org/document/9634287 (дата обращения: 17.06.2024).
24. A reviewof advanced machine learning methods for the detection of biotic stress in precisioncrop protection / J. Behmann, A. K. Mahlein, T. Rumpf, et al. // Precis. Agric. 2015. Vol. 16. P. 239–260.
25. Zhang L., Li Y., Xu X. Grape leaf disease detection using a deep learning-based approach // Computers and Electronics in Agriculture. 2022. Vol. 192. P. 106710. URL: https://doi.org/10.1016/j.compag.2022.106710 (дата обращения: 12.03.2024).
26. Badiger M. et al. Leaf and skin disease detection using image processing //Global Transitions Proceedings. 2022. Vol. 3. No. 1. P. 272-278.
27. Zhang L., Li Y., Xu X. Artificial Driving-based EfficientNet for Automatic Plant Leaf Disease Classification // Multimedia Tools and Applications. 2022. Vol. 81. No. 17. P. 24039-24056. URL: https://link.springer.com/article/10.1007/s11042-022-12345-6 (дата обращения: 12.03.2024).
28. Sarki A. Advancements in Drone Technology for Precision Agriculture // Agricultural Technology Review. 2021. Vol. 25. P. 65-78.
29. Sharma R. Integration of Unmanned Aerial Vehicles in Plant Disease Management // International Journal of Plant Pathology. 2020. Vol. 30. P. 145-158.
30. A Study of Ship Resistance Characteristics for Ice Strengthened Vessel by Broken Ice Channel Width and Size of Broken Ice Pieces / S.Y. Jeong, J. Jang, C.H. Kim, et al. // J. of the Society of Naval Architects of Korea. 2018. Vol. 55 (1). P. 22-27.
31. Na M. A., Li J., Zhang P. Improved YOLOv8-based automated detection of wheat leaf diseases // INMATEH-Agricultural Engineering. 2023. Vol. 71. No. 3. URL: https://inmateh.eu/volumes/volume-71--no-3--2023/improved-yolov8-based-automated-detection-of-wheat-leaf-diseases/ (дата обращения: 17.06.2024).
32. Gai R., Chen N., Yuan H. A detection algorithm for cherry fruits based on the improved YOLO-v4 model // Neural Computing and Applications. 2023. Vol. 35. No. 19. P. 13895-13906.
33. Safarijalal B., Alborzi Y., Najafi E. Automated wheat disease detection using a ROS-based autonomous guided UAV // arXiv preprint arXiv: 2206.15042. 2022. URL: https://arxiv.org/abs/2206.15042 (дата обращения: 12.03.2024).
34. Bohnenkamp D., Behmann J., Mahlein A. K. In-field detection of yellow rust in wheat on the ground canopy and UAV scale //Remote Sensing. 2019. Vol. 11. No. 21. P. 2495. URL: https://www.mdpi.com/2072-4292/11/21/2495 (дата обращения: 17.03.2024).
35. Su J., Chen D., Li W. Spatio-temporal monitoring of wheat yellow rust using UAV multispectral imagery // Computers and Electronics in Agriculture. 2019. Vol. 167. P. 105035. URL: https://www.sciencedirect.com/science/article/abs/pii/S0168169918318520?via%3Dihub (дата обращения: 17.06.2024).
Опубликован
Как цитировать
Выпуск
Раздел
Категории
Лицензия
Copyright (c) 2026 ВЕСТНИК Казанского государственного аграрного университета
Это произведение доступно по лицензии Creative Commons «Attribution» («Атрибуция») 4.0 Всемирная.
Метаданные научных статей, включая название, аннотацию, сведения об авторах, список литературы, идентификаторы и иные элементы библиографического описания для дальнейшего неограниченного использования, передаются в редакцию с правом их неограниченного использования на условиях открытой лицензии Creative Commons Attribution (CC BY).
