Plant Leaf Disease Detection Using Mask R-CNN

1. Executive Summary

The Plant Leaf Disease Detection project leverages deep learning techniques, specifically Mask R-CNN and U-Net architectures, to detect and classify plant leaf diseases with high accuracy. The system performs real-time segmentation and classification, enabling early identification of diseases for better crop management. The application is deployed on a Flask-based web platform with a responsive HTML/CSS/JavaScript front end, making it easily accessible to farmers and agricultural researchers.

Key Achievements:

Developed and trained Mask R-CNN and U-Net models for precise segmentation and classification
Achieved high accuracy in identifying multiple plant diseases across different species
Implemented a user-friendly web interface for real-time image analysis
Enabled disease management decision-making through fast detection results
Completed full development within a 10-day timeline

2. Project Scope & Features

2.1 Core Functionality

Primary Features:

Real-time plant leaf disease detection from uploaded images
Deep learning-based segmentation using Mask R-CNN and U-Net
High-accuracy classification of multiple disease types
Web deployment for accessible agricultural use
Responsive and intuitive front-end interface

2.2 Supported Plant Species & Diseases

Multiple plant species including tomato, potato, and grape
Diseases such as early blight, late blight, leaf mold, and more
System designed to be scalable for new species and diseases

3. Technology Stack

3.1 Core Technologies

Python
Primary programming language

Flask
Web application framework

Mask R-CNN
Instance segmentation model

U-Net
Semantic segmentation model

NumPy
Numerical computing and array operations

OpenCV
Image processing library

HTML/CSS/JavaScript
Front-end development

4. Project Statistics

10

Days to Complete

95%+

Model Accuracy

Multiple

Supported Species

Real-time

Detection Speed

5. Conclusion

The Plant Leaf Disease Detection system demonstrates the effectiveness of deep learning in agricultural problem-solving. By integrating Mask R-CNN and U-Net models into an accessible web application, the project delivers a practical tool for early disease detection, helping farmers and researchers make timely decisions to protect crops and maximize yield potential.