AUTOMATED IDENTIFICATION OF ACTIVATED SLUDGE BIOINDICATORS BASED ON THE YOLOV8 NETWORK MODEL
DOI:
https://doi.org/10.32689/maup.it.2025.4.30Keywords:
object recognition, automatic image analysis, deep learning, neural network, YOLOv8, bioindicators, activated sludgeAbstract
The paper presents the results of a study focused on the development and training of a YOLOv8 deep learning neural network for the automated recognition of bioindicators in microscopic images of activated sludge. The aim of the research is to develop and test a model for automatic recognition of activated sludge bioindicators to improve the efficiency of real-time monitoring of the biological wastewater treatment process. The main task is to create an intelligent tool capable of identifying morphological objects (bioindicators) in activated sludge images, enabling rapid assessment of bioprocess performance and adaptive control of treatment parameters. Methodology. The study employs the modern YOLOv8 architecture, which combines high processing speed, accuracy, and the ability to train on relatively small datasets. The model was trained on labeled samples of microscopic images containing two types of activated sludge bioindicators. The training results were validated using a separate dataset, and the model’s stability was evaluated under varying lighting and contrast conditions. Scientific novelty. For the first time, the applicability of the YOLOv8 architecture to the task of automatic recognition of activated sludge bioindicators has been demonstrated. The model showed the ability to accurately detect most bioindicators and maintained stable performance even under complex image conditions. This approach opens up new possibilities for the development of intelligent monitoring systems for biological treatment processes, capable of self-adaptation and real-time decision-making. Conclusions. The conducted study confirmed the effectiveness of using YOLOv8 for the identification of activated sludge bioindicators. The obtained results demonstrate satisfactory model accuracy at the demonstration training stage, while further dataset expansion is expected to improve system precision and reliability for industrial applications. The proposed approach can serve as the basis for the development of automated systems for monitoring activated sludge condition and controlling biological wastewater treatment processes.
References
Забара І. І. Саморегуляція активного мулу. Екологічна безпека: проблеми і шляхи вирішення: зб. наук. статей XVІІІ Міжнародної науково-практичної конференції (м. Харків, 15–16 вересня 2022 р.) / УКРНДІЕП. Харків, 2022. С. 167–174.
Матвійчук М. Оптимізація стічних вод за допомогою нейронної мережі при очищенні промислових стічних вод. Рівне: НУВХП. 2023. Т. 28, № 2. С. 107–112. DOI: 10.15407/jai2023.02.107.
Alekseevsky D. G., Chernysh Ye. Yu., Shtepa V. N. Formalization of the task of creating a mathematical model of combined wastewater treatment processes. Journal of Engineering Sciences. 2021. Vol. 8, No. 2. P. H1–H7. DOI: 10.21272/jes.2021.8(2).h1.
Chub I., Airapetian T., Karahiaur A., Zabara I. The use of biological activation of microorganisms of activated sludge to increase the efficiency of wastewater treatment. AIP Conference Proceedings. 2023. Vol. 2490, Issue 1. Art. no. 060033. DOI: 10.1063/5.0123327.
Dziadosz M., Majerek D., Łagód G. Microscopic Studies of Activated Sludge Supported by Automatic Image Analysis Based on Deep Learning Neural Networks. Journal of Ecological Engineering. 2024. Vol. 25, No. 4. P. 360–369. DOI: 10.12911/22998993/185317.
Haluza A., Shtepa V., Yushenko V. Development of control and forecasting system for the effective functioning of biological wastewater treatment plants in the city of Vitebsk based on neural networks. Herald of Polotsk State University. Series F: Civil Engineering. Applied Sciences. 2023. No. 35(3). P. 2–10. DOI: 10.52928/2070-1683-2023-35-3-2-10.
How to install YOLOv8. URL: https://yolov8.org/how-to-install-yolov8/ (дата звернення: 19.10.2025).
Image Class Selector. URL: https://github.com/buchslava/image-class-selector (дата звернення: 19.10.2025).
Moldovan A., Nuca I. Automation of Wastewater Treatment Plant. 2019 International Conference on Electromechanical and Energy Systems (SIELMEN), 9–11 October 2019, Craiova, Romania. IEEE, 2019. DOI: 10.1109/SIELMEN.2019.8905867.
Redmon J. You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016. P. 779–788. DOI: 10.1109/CVPR.2016.91.
Singh N. D., Murugamani C., Kshirsagar P., Tirth V., Islam S., Qaiyum S., Suneela B., Duhayyim M.A., Waji Y.A. IOT based smart wastewater treatment model for industry 4.0 using artificial intelligence. Scientific Programming. 2022. Article ID 5134013. DOI: 10.1155/2022/5134013.
Stawarczyk M., Stawarczyk K. Use of the ImageJ program to assess the damage of plants by snails. Chemistry-Didactics-Ecology-Metrology. 2015. Vol. 20, No. 1–2.
Tarafdar A., Shahi N. Ch., Singh A., Sirohi R. Artificial neural network modeling of water activity: a low energy approach to freeze drying. Food and Bioprocess Technology. 2018. Vol. 11, No. 1. P. 164–171. DOI: 10.1007/s11947-017-2002-4.
Wang Z., Dai H., Chen B., Cheng S., Sun Y., Zhao J., Guo Z., Cai X., Wang X., Li B., Geng H. Effluent quality prediction of the sewage treatment based on a hybrid neural network model: Comparison and application. Journal of Environmental Management. 2024. Vol. 351. Art. no. 119900. DOI: 10.1016/j.jenvman.2023.119900.
Xinyi Q. The application of artificial intelligence – artificial neural networks – in wastewater treatment. E3S Web of Conferences. 2023. Vol. 393. Art. no. 03003. DOI: 10.1051/e3sconf/202339303003.
Yang L., Honarvar Shakibaei Asli B. MSConv-YOLO: An Improved Small Target Detection Algorithm Based on YOLOv8. Journal of Imaging. 2025. Vol. 11, No. 8. Art. no. 285. DOI: 10.3390/jimaging11080285.
