Plant Disease Detection using CNN

This repository contains an implementation of plant disease detection using a Convolutional Neural Network (CNN) built from scratch. The repository also contains a web app allowing users to upload an image of a plant and the CNN model predicts whether the plant is healthy or if it is affected by powdery mildew or rust.

---

Few Important Links:

• Model Weights:
  • https://www.mediafire.com/file/1j3dwd2pzs7xc13/weights.rar/file.
• Deployed App Link:
  • https://plant-disease-detection-pkue.onrender.com/
• Model API Link:
  • https://arpy8-plant-detection-api.hf.space/docs

---

Table of Contents

• Introduction
• Prerequisites
• Usage
• File Structure
• Model
• License

Introduction

Plant diseases can have a detrimental impact on crop yield. Early detection and intervention are crucial to prevent the spread of diseases and minimize damage. The Streamlit web application provides an easy-to-use interface for users to upload plant images, which are then processed using a CNN model to predict the health status of the plant.

Prerequisites

Before running the application, make sure you have the following dependencies installed:

• Python 3.x
• TensorFlow
• Streamlit
• Numpy
• Pillow (PIL)
• The model weights file (plant_disease_classifier.h5) should be available in the weights directory.

You can install the required packages using pip:

pip install tensorflow streamlit numpy pillow

Usage

1. Clone the repository:

   git clone https://github.com/arpy8/Plant_Disease_Detection.git
   cd Plant_Disease_Detection_using_CNN

2. Place the model weights file (plant_disease_classifier.h5) in the weights directory.

3. Run the Streamlit app:

   streamlit run app.py

4. The web app will open in your default web browser, allowing you to upload plant images for disease classification.

File Structure

├── LICENSE
├── README.md          <- The top-level README for developers/collaborators using this project.
├── app.py             <- Source Code for the streamlit web app hosted by me.
│
└── assets             <- Folder containing the assets for the web app
│   ├── bg.png         <- Background image
│   ├── logo.png       <- Logo image
│   └── sample         <- Contains random image from the dataset for the web app
│
│
├── notebook           <- Folder containing the notebook used for training the CNN model
│   └── Plant-Disease-Detection-using-CNN.ipynb      <- Contains the model weights
│
└── weights            <- Folder containing the dataset reports/results of this project
    └── plant_disease_classifier.h5                  <- Model weights of the cnn model

Model

The model used for plant disease classification is a CNN model. It has been trained to classify plants into one of the following categories:

• Healthy
• Powdery
• Rust

The model's architecture and weights are loaded from the plant_disease_classifier.h5 file.

License

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