app.py

import streamlit as st
import cv2
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.image import img_to_array
from PIL import Image

# Load the pre-trained model
model = load_model('plant_diseases.h5')

# Class labels (replace with your own classes)
class_labels = [
    'Piment: Bacterial_spot', 
    'Piment: healthy', 
    'Pomme de terre: Early_blight', 
    'Pomme de terre: Late_blight', 
    'Pomme de terre: Healthy', 
    'Tomate: Bacterial Spot', 
    'Tomate: Early Blight', 
    'Tomate: Late Blight', 
    'Tomate: Leaf mold', 
    'Tomate: Septoria leaf spot', 
    'Tomate: Spider mites', 
    'Tomate: Spot', 
    'Tomate: Yellow Leaf Curl', 
    'Tomate: Virus Mosaïque', 
    'Tomate: Healthy'
]

def preprocess_image(image, image_size=(224, 224)):
    # Convert image to grayscale
    image = np.array(image.convert('L'))
    # Resize image
    image = cv2.resize(image, image_size)
    # Prepare image for the model
    image = img_to_array(image)
    image /= 255.0
    image = np.expand_dims(image, axis=0)
    return image

# Streamlit app setup
st.title("Classification des Maladies des Plantes")
st.write("Téléchargez une image de plante pour la classification")

uploaded_file = st.file_uploader("Choisissez une image...", type=["jpg", "jpeg", "png"])

if uploaded_file is not None:
    # Display the uploaded image
    image = Image.open(uploaded_file)
    st.image(image, caption='Image téléchargée', use_column_width=True)

    st.write("Classification en cours...")

    # Preprocess the image
    processed_image = preprocess_image(image)

    # Make predictions
    predictions = model.predict(processed_image)
    probabilities = predictions[0]

    # Display probabilities for each class
    for i, label in enumerate(class_labels):
        if probabilities[i] > 0:
            st.write(f"{label}: {probabilities[i]:.2f}")

    # Show predicted class
    predicted_class = class_labels[np.argmax(probabilities)]
    st.write(f"Classe prédite: {predicted_class}")