Pimple et al. (2025) AI-Driven Agriculture Monitoring System

Identification

• Journal: INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT
• Year: 2025
• Date: 2025-12-11
• Authors: Santosh Tulshiram Pimple, Shravani Kishor Hore
• DOI: 10.55041/ijsrem55131

Research Groups

Santosh Tulshiram Pimple and Shravani Kishor Hore from Prof. Ramkrishna More College, Pradhikaran, Pune, India.

Short Summary

This paper presents an AI-driven agriculture monitoring system that integrates real-time crop disease detection, soil moisture prediction, automated irrigation, and weather-risk alerts. The system leverages machine learning and sensor data to provide farmers with precise, timely insights, aiming to enhance crop health, productivity, and agricultural sustainability.

Objective

• To design, develop, test, and validate an intelligent, integrated AI-Driven Agriculture Monitoring System that uses machine learning models, sensors, and predictive analytics to detect crop diseases early, estimate soil moisture levels, automate irrigation, and send timely weather alerts, thereby enhancing productivity, reducing resource wastage, increasing crop health, and supporting sustainable farming.

Study Configuration

• Spatial Scale: The system is designed for modern farming applications, with a case study mentioning a mid-size vegetable farm (approximately 16.19 hectares). It aims for region-specific models and addresses challenges of localization for diverse crops and climates.
• Temporal Scale: Real-time data analysis, continuous monitoring, short-term forecasting (soil moisture), timely alerts, and automated decision-making.

Methodology and Data

• Models used: Machine learning models, Deep learning (Convolutional Neural Networks - CNNs, transfer learning, lightweight models for edge deployment), Long Short-Term Memory (LSTM) for time-series soil moisture prediction, Random Forests, Gradient Boosting, YOLO, Vision Transformers.
• Data sources: In-situ sensors (capacitive/YL-69 soil moisture sensors, DS18B20 soil temperature sensors, pH & EC sensors, DHT22, BMP180, rain gauge, anemometer weather sensors, BH1750, ML8511 light/UV sensors, LWS-mk2 leaf wetness sensors), remote sensing (satellite microwave/optical indices), image data (plant images via mobile/drone), environmental parameters, historical and real-time data, farmer reports.

Main Results

• The system achieves 90–95% accuracy for crop disease detection using deep learning models.
• It provides real-time insights, accurate predictions (soil moisture, disease outbreaks, rainfall, yield), and automated decision-making capabilities.
• Key benefits include reduced manual labor, minimized crop loss risk, optimized water usage, and improved agricultural efficiency and sustainability.
• The system supports automated design generation, accelerating the creation of system architecture, UI layouts, data pipelines, and component configurations.
• It enhances accuracy in crop health monitoring and provides predictive analytics for effective farm management.

Contributions

• Development of an intelligent, integrated AI-driven platform combining multiple critical agricultural monitoring functions (disease detection, soil moisture prediction, automated irrigation, weather alerts) into a single, cohesive system.
• Application of diverse machine learning models (CNNs, LSTM, etc.) for real-time analysis and predictive capabilities in agriculture.
• Emphasis on automated design generation and a modular codebase to ensure reproducibility and scalability of the system.
• Addresses challenges of traditional farming by offering data-driven, precise, and automated solutions, leading to improved resource efficiency and sustainability.

Funding

No explicit funding information is provided in the paper.

Citation

@article{Pimple2025AIDriven,
  author = {Pimple, Santosh Tulshiram and Hore, Shravani Kishor},
  title = {AI-Driven Agriculture Monitoring System},
  journal = {INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT},
  year = {2025},
  doi = {10.55041/ijsrem55131},
  url = {https://doi.org/10.55041/ijsrem55131}
}