Plant Disease Detection Using Deep Learning
A deep-learning based system that detects crop diseases from leaf images using a Convolutional Neural Network (CNN). The project includes:
A trained model (.h5)
Prediction script
Remedies & disease descriptions
A Streamlit web app
Example images and screenshots
🚀 Features
✔ Classifies plant diseases from leaf images ✔ Shows the confidence score ✔ Displays description + remedies for the predicted disease ✔ Simple Streamlit web interface ✔ Lightweight, fast, and easy to run
📦 Installation & Setup
1️⃣ Clone the repository
git clone https://github.com/YOUR-USERNAME/YOUR-REPO-NAME.git cd YOUR-REPO-NAME
2️⃣ Install dependencies
pip install -r requirements.txt
3️⃣ Download the trained model
Download the model from Drive:
👉 Model Link: https://drive.google.com/file/d/190XKlIX47r497C2W54c1Env5c-WAH8zu/view?usp=sharing
Place the file in the root folder and rename it:
best_model.h5
streamlit run app.py
This opens the web UI where you can upload a leaf image, and the model will display:
Predicted disease
Confidence
Description
Remedies
🧪 Example Output
Here’s a sample CLI run:
python predict.py sample_images/sample1.jpg
Output:
Predicted: Tomato___Late_blight (92.17%)
🖼 App Screenshots
🏠 Home Page
📤 Upload Page
📊 Output Page
📁 Project Structure
📂 Plant Disease Detector │── app.py │── predict.py │── requirements.txt │── class_indices.npy │── remedies.json │── README.md │── sample_images/ │── screenshots/
🧠 Model Training Info
Framework: TensorFlow / Keras
Input size: 224 × 224
Trained on: PlantVillage dataset
Optimizer: Adam
Loss: Categorical Crossentropy
🙌 Acknowledgements
This project was created for academic purposes and is based on open datasets and deep learning frameworks.