🥔Potato Plant Disease Detection 🩺

This repository contains the source code and documentation for a web-based application that helps farmers detect potato plant diseases by analyzing leaf images. Users can either upload a photo from their gallery or capture one using their device's camera. The project integrates machine learning for accurate disease prediction and is deployed using modern frameworks and tools.

Purpose💡

Agriculture faces significant challenges from plant diseases, which can lead to reduced yields and economic loss. By providing an easy-to-use tool for farmers, this project aims to:

• Quickly identify potato plant diseases.
• Reduce dependence on manual diagnosis.
• Provide insights for timely treatment and disease prevention.

Features✨

• Image Upload and Capture: Upload images from the gallery or capture directly via the camera.
• Machine Learning Model: Built using TensorFlow and Keras with 93.88% accuracy.
• Visualization: Data insights visualized using matplotlib and seaborn.
• API Integration: Backend powered by FastAPI with TensorFlow Serve.
• Cloud Deployment: TensorFlow Lite model deployed on Google Cloud Platform (GCP).
• API Testing: Endpoints tested using Postman.

Tech Stack

Frontend

• Framework: React.js
• Key Features: Responsive design, user-friendly interface for both desktop and mobile devices.

Backend

• Framework: FastAPI
• Model Serving: TensorFlow Serve for real-time predictions.
• Hosting: Deployed on a local server or cloud environments.

Machine Learning

• Libraries: TensorFlow, Keras, seaborn, matplotlib
• Development Platform: Google Colab for efficient training and experimentation.
• Dataset: Sourced from Kaggle, specifically curated for potato plant diseases.

Techniques Used:

• Data Augmentation: To enhance dataset diversity and improve model robustness.
• Transfer Learning: Leveraging pre-trained models for faster and more accurate training.

Deployment

• Cloud Service: Google Cloud Platform (GCP)
• API Testing: Postman for ensuring reliable and secure API endpoints.

Dataset and Model

• Source: Kaggle https://www.kaggle.com/datasets/arjuntejaswi/plant-village where I only took potato plant leaves datasets omit other leaves datasets
• Images: Labeled images of potato leaves, categorized by disease types like 'Potato Early blight', 'Potato Late blight' and 'healthy leaves'.

Model Training

• Architecture: CNN (Convolutional Neural Network) based on TensorFlow and Keras.
• Performance: Achieved 93.88% accuracy on the validation dataset.
• Visualization: Model accuracy and loss graphs plotted using matplotlib.

Installation and Setup

Prerequisites

• Python 3.9+
• Node.js and npm
• Google Cloud Platform (GCP) account
• Postman for API testing

Clone the Repository

git clone https://github.com/fasinfasi/Potato_Disease_Prediction.git
cd potato-disease-detection  

1. Frontend Setup

Navigate to the frontend directory:

cd frontend

Install dependencies:

npm install

Run the React development server:

npm start

2. Backend Setup

Navigate to the backend directory:

cd api

Start the FastAPI server:

uvicorn main:app --reload

Usage

1. Access the Web App

Open the frontend in your browser by navigating to http://localhost:3000 (default React port).

2. Upload or Capture Image

Select or capture an image of a potato leaf. Submit the image for disease detection.

3. View Results

The model predicts the type of disease or confirms a healthy leaf. Results are displayed on the screen.

Feel free to use and adapt this README! Let me know if you'd like further refinements🤗.