Automated Spray Control Based on Severity Level for Tomato Plants using Deep Learning and Image Processing

Overview

The rapid and accurate identification of plant diseases is crucial for maintaining crop health and agricultural productivity. Traditional methods are often manual, time-consuming, and prone to errors. In recent years, advancements in machine vision technology, particularly leveraging deep learning and image processing techniques, have shown promise in addressing these challenges.

This repository contains code and resources for an end-to-end system designed to automate the detection and severity estimation of diseases in tomato plants. Our approach combines state-of-the-art deep learning models with image processing algorithms to analyze digital images of tomato leaves. This allows for not only detecting the presence of disease but also quantifying its severity.

Key Features

• Automated Disease Detection: Utilizes deep learning models for accurate identification of disease presence in tomato plants.
• Severity Estimation: Provides quantitative assessment of disease severity, crucial for effective disease management strategies.
• Integration with Agricultural Practices: Designed to be integrated into existing agricultural workflows, facilitating informed decision-making and optimized treatment plans.
• Cost Savings: Enables significant reduction in pesticide usage by precisely measuring disease severity and implementing targeted treatment strategies.

Repository Structure

• Disease Detection/: Contains dataset for training and implementation of the deep learning models for Disease Detection in Tomato Plants.
• Segmentation/: Contains code and dataset for image segmentation related to disease detection.
• Severity Estimation/: Includes algorithms and resources for quantifying disease severity from detected regions.
• Validation/: Stores the output and results from experiments and validation tests.

Getting Started

To get started with using or contributing to this repository, follow these steps:

1. Clone the Repository:

   git clone https://github.com/Neural-Ninja/Crop-Disease-Detection-and-Severity-Estimation-PLOS-ONE.git
   cd Crop-Disease-Detection-and-Severity-Estimation-PLOS-ONE
   

2. Set up environment:

• Install required dependencies as listed in requirements.txt.
• Configure any environment variables or paths needed for your setup.

3. Run the code:

• Execute the scripts/notebooks in the respective directory for disease detection, segmentation and severity estimation.

4. Explore the documentation:

• Refer to Docs/ for detailed documentation on the algorithms used, data formats, and experiment results.

How to Contribute

Contributions to improve the functionality or performance of this system are welcome. Here are some ways you can contribute:

• Implement new features or improvements.
• Test the system with different datasets or scenarios.
• Report bugs or issues you encounter.
• Provide feedback on usability and documentation.

Please follow our contribution guidelines and code of conduct when contributing to this project.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Note

This repository is made for our research publication under PLOS-One Journal.