app.py

import streamlit as st
from ultralytics import YOLO
import cv2
from gtts import gTTS
import os
import tempfile

# Load YOLOv8 model
@st.cache_resource
def load_model():
    return YOLO('yolov8n.pt')  # Automatically downloads YOLOv8 pre-trained model

model = load_model()

# Streamlit app title
st.title("Real-Time Object Detection for Blind Assistance")
st.write("This application detects objects in real-time from a webcam feed and provides audio feedback.")

# Start detection button
start_detection = st.button("Start Real-Time Detection")

if start_detection:
    # Open webcam
    st.write("Starting webcam... Press 'q' to stop.")
    cap = cv2.VideoCapture(0)  # 0 for default webcam, change if multiple webcams are connected

    # Check if webcam is opened
    if not cap.isOpened():
        st.write("Error: Could not open webcam.")
    else:
        stframe = st.empty()  # Placeholder for video frames

        while cap.isOpened():
            ret, frame = cap.read()
            if not ret:
                break

            # Perform object detection
            results = model(frame)

            # Get detected object names
            detected_objects = [model.names[int(box.cls)] for box in results[0].boxes]

            # Generate audio feedback
            if detected_objects:
                objects_text = ", ".join(set(detected_objects))
                summary_text = f"Detected: {objects_text}."
                st.write(summary_text)  # Display detected objects in Streamlit

                # Convert text to speech with gTTS
                tts = gTTS(text=summary_text, lang='en')
                audio_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
                tts.save(audio_file.name)
                os.system(f"mpg123 {audio_file.name}")  # Play the audio summary

            # Annotate frame with bounding boxes
            annotated_frame = results[0].plot()

            # Display the annotated frame in Streamlit
            stframe.image(annotated_frame, channels="BGR", use_column_width=True)

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

        # Release webcam and close OpenCV windows
        cap.release()
        cv2.destroyAllWindows()
        st.write("Real-Time Detection Stopped.")