Crop Harvesting using Machine Learning (ML) and Internet of Things (IOT)
- Version
- Download 17
- File Size 378.22 KB
- File Count 1
- Create Date 11 April 2025
- Last Updated 11 April 2025
Crop Harvesting using Machine Learning (ML) and Internet of Things (IOT)
Authors:
| Arunabh Sharma | Theertha K Sunil | Jay Prakash Sah |
| Computer Science & Engineering | Computer Science & Engineering | Computer Science & Engineering |
| Jain (Deemed to be University) | Jain (Deemed to be University) | Jain (Deemed to be University) |
| Bangalore, Karnataka, India | Bangalore, Karnataka, India | Bangalore, Karnataka, India |
| 21btrcs263@jainuniversity.ac.in | 21bttcn001@jainuniversity.ac.in | 21btrcs244@jainuniversity.ac.in |
| Abishek Kumar Sah | Shiv Ranjan Kumar Adhikari | Dr. Rhea Sriniwas |
| Computer Science & Engineering | Computer Science & Engineering | Computer Science & Engineering |
| Jain (Deemed to be University) | Jain (Deemed to be University) | Jain(Deemed to be University) |
| Bangalore, Karnataka, India | Bangalore, Karnataka, India | Bangalore, Karnataka, India |
| 21btrcs256@jainuniversity.ac.in | 21btrcs230@jainuniversity.ac.in | rhea.sriniwas@jainuniversity.ac.in |
Abstract- Accurate crop prediction is essential for optimizing agricultural productivity and ensuring food security, particularly in regions where farming decisions are heavily influenced by soil and climatic conditions. Traditional crop recommendation systems often lack real-time adaptability and the ability to integrate diverse data sources such as soil sensors and live weather feeds. This project introduces a real-time crop prediction system that leverages a Support Vector Machine (SVM) classifier for precise and reliable crop recommendations. The system integrates heterogeneous data inputs, including soil nutrient levels (N, P, K), pH value obtained from an Arduino-based sensor, and dynamic weather parameters such as temperature, humidity, and rainfall retrieved via public APIs.
A Flask-based web interface facilitates seamless user interaction, automatically collecting environmental data to reduce manual input and improve usability. The SVM model, trained on the 'Crop_recommendation.csv' dataset, maps real-time data to the most suitable crops with approximately 95% accuracy. Label encoding is used for effective handling of categorical crop labels.
Evaluation results indicate that the system delivers accurate predictions with low latency, confirming its potential in real-world agricultural settings. By combining sensor data, geolocation-based weather insights, and machine learning, the proposed system exemplifies a scalable solution for smart farming, paving the way for sustainable and efficient agricultural practices.
Keywords— Crop Prediction, Support Vector Machine(SVM), Real-time agriculture system, flask web interface, sensor-based soil analysis, weather API integration, precision farming, smart agriculture, low-latency prediction, sustainable farming solutions.