realtime_detection.py

import cv2
import numpy as np
from tensorflow.keras.models import load_model
from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
import os

# Load the trained model
model_path = "plant_classifier_transfer_model.h5"  # Replace with the path to your saved model
model = load_model(model_path)

# Load class names from the training directory
class_labels = sorted(os.listdir(r"C:\Users\Ramachandra\OneDrive\Desktop\saad\otrain"))  # Path to your train directory

# Parameters
img_height, img_width = 224, 224  # Image dimensions used during training

# Initialize the video capture (0 for the default laptop camera)
cap = cv2.VideoCapture(0)

def predict_frame(frame):
    """
    Preprocesses the frame and predicts the class using the trained model.
    Args:
        frame: A single frame captured from the video feed.
    Returns:
        str: Predicted class label.
    """
    # Resize the frame to the input size of the model
    resized_frame = cv2.resize(frame, (img_width, img_height))
    
    # Preprocess the frame
    preprocessed_frame = preprocess_input(resized_frame)
    preprocessed_frame = np.expand_dims(preprocessed_frame, axis=0)  # Add batch dimension
    
    # Predict the class
    predictions = model.predict(preprocessed_frame)
    predicted_class_index = np.argmax(predictions)
    predicted_class_label = class_labels[predicted_class_index]
    
    return predicted_class_label

print("Press 'q' to quit the application.")

# Real-Time Object Detection
while True:
    ret, frame = cap.read()  # Capture a frame from the camera
    if not ret:
        break

    # Flip the frame horizontally for natural mirroring
    frame = cv2.flip(frame, 1)
    
    # Predict the class
    predicted_label = predict_frame(frame)

    # Display the predicted label on the frame
    cv2.putText(frame, f"Prediction: {predicted_label}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 
                1, (0, 255, 0), 2, cv2.LINE_AA)
    
    # Show the frame in a window
    cv2.imshow("Real-Time Plant Detection", frame)

    # Print the predicted label to the console
    print(f"Predicted Class: {predicted_label}")

    # Break the loop if 'q' is pressed
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

# Release the video capture and close all OpenCV windows
cap.release()
cv2.destroyAllWindows()