IoT Based Real Time Crop Recommendation System Using Random Forest Classifier

Agriculture continues to be a pillar of worldwide food security and economic stability but farmers often face challenges making an appropriate crop selection as environmental conditions constantly shift. This research presents an adaptive crop suggestion system that combines IoT-based field sensors...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2025 International Conference on Networks & Advances in Computational Technologies (NetACT) S. 1 - 6
Hauptverfasser: Kannan, Anjana, Tony, Aneena Edakkalathur, Vinoj, Ansen, Faimon, Siyona, Moozhippurath, Bineesh, Mohandas, Arjun Kizhupadath
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 07.08.2025
Schlagworte:
Agriculture Yields
Crop Recommendation
Crops
Feeds
Humidity
Machine Learning
Predictive models
Random Forest
Random forests
Real-time systems
Recommender systems
Sensor systems
Soil
Soil Behavior Analysis
Temperature sensors
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Agriculture continues to be a pillar of worldwide food security and economic stability but farmers often face challenges making an appropriate crop selection as environmental conditions constantly shift. This research presents an adaptive crop suggestion system that combines IoT-based field sensors with machine learning to support more informed decision-making in precision agriculture. As opposed to conventional methods based on stationary soil data, our system continuously evaluates real-time environmental factors-temperature, humidity, and precipitation (provided through OpenWeatherMap API)-in addition to IoT-measured soil factors like NPK content and pH. Using a trained Random Forest model with a large crop dataset, the system reaches a prediction accuracy of 98.9%, well above the performance of traditional methods. The hardware design integrates NPK and pH sensors with an Arduino microcontroller for unconstrained data capture, while the software infrastructure includes a Flask-driven backend and a user-friendly React-based interface for farmer usability. By integrating real-time sensor feeds with strong machine learning, our architecture provides actionable crop recommendations, minimizing resource wastages, optimizing yields, and promoting sustainable agriculture. This project fills key shortcomings in static crop models, presenting a scalable and adaptive tool for contemporary agronomy.
DOI:10.1109/NetACT65906.2025.11188137