EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

IoT and Machine Learning-Based Smart Agriculture Monitoring System

Devendra Prasad, Dinesh Verma, Navneet, Rashmi

International Journal of Research and Review in Applied Science, Humanities, and Technology · 28.12.2024 · 252-262

DOI: 10.71143/hqyca248 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Research and Review in Applied Science, Humanities, and Technology, online erschienen am 28.12.2024. Verfasst von Devendra Prasad und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Irrigation timing and crop selection strongly influence farm productivity, yet on most small farms both decisions are taken without field data. This paper presents an Internet of Things (IoT) and machine learning (ML) based smart agriculture monitoring system built on a four-layer architecture: a low-cost ESP32 sensor node that measures soil moisture, air temperature, humidity, soil pH, soil nutrients (N, P, K) and rain; MQTT communication; a cloud analytics layer; and a farmer dashboard. The system provides ML-based crop recommendation, sensor-driven irrigation control and short-term soil-moisture forecasting. Seven classifiers were evaluated on a public dataset of 2,200 samples and 22 crops. Gaussian Naive Bayes and Random Forest reached the highest test accuracy of 99.55%, and Naive Bayes was selected for deployment because it gave the best cross-validation accuracy (99.49%) with a model size of only 3.4 kB. Its accuracy stays above 97% when noise equal to 10% of each feature’s standard deviation is added, although a support vector machine is more robust under heavier noise. The sensing and irrigation functions were evaluated in a 60-day software simulation with an emulated sensor node. Sensor-based control reduced irrigation water by 26.5% compared with a timer-based schedule (30.0 ± 6.8% over 30 simulated seasons) without causing crop water stress, and an XGBoost model forecast soil moisture six hours ahead with a root-mean-square error of about 1.0% volumetric water content.

Zitieren

APA

Prasad, D., Verma, D., Navneet, & Rashmi (2024). IoT and Machine Learning-Based Smart Agriculture Monitoring System. International Journal of Research and Review in Applied Science, Humanities, and Technology, 252-262. https://doi.org/10.71143/hqyca248

Vancouver

Prasad D, Verma D, Navneet, Rashmi. IoT and Machine Learning-Based Smart Agriculture Monitoring System. International Journal of Research and Review in Applied Science, Humanities, and Technology. 2024:252-262. doi:10.71143/hqyca248

BibTeX
@article{Prasad2024,
  author = {Prasad, Devendra and Verma, Dinesh and Navneet and Rashmi},
  title = {IoT and Machine Learning-Based Smart Agriculture Monitoring System},
  journal = {International Journal of Research and Review in Applied Science, Humanities, and Technology},
  year = {2024},
  pages = {252--262},
  doi = {10.71143/hqyca248}
}
Kostenlos starten →