File size: 4,612 Bytes
5a75ffb
58b9573
 
5a75ffb
 
 
58b9573
b40700a
5a75ffb
58b9573
5a75ffb
 
 
 
 
b40700a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5a75ffb
b40700a
 
5a75ffb
770a476
5a75ffb
b40700a
 
 
 
 
 
 
 
5a75ffb
b40700a
 
5a75ffb
b40700a
5a75ffb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b40700a
 
5a75ffb
 
 
b40700a
 
 
5a75ffb
 
 
 
 
 
 
 
b40700a
5a75ffb
 
 
b40700a
 
5a75ffb
 
 
 
b40700a
 
 
 
 
 
 
 
 
5a75ffb
 
 
b40700a
 
58b9573
b40700a
5a75ffb
b40700a
ce1306d
5a75ffb
b40700a
ce1306d
b40700a
5a75ffb
 
b40700a
 
 
 
5a75ffb
 
ce1306d
5a75ffb
ce1306d
5a75ffb
 
 
770a476
 
b40700a
ce1306d
5a75ffb
b40700a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
import os
import streamlit as st
from ultralytics import YOLO
import cv2
import random
import time
from gtts import gTTS
from playsound import playsound
from datetime import datetime, timedelta

# Load YOLOv8 model
yolo = YOLO("yolov8n.pt")

# Streamlit app layout
st.set_page_config(page_title="Assistive Vision App", layout="wide")
st.markdown(
    """
    <style>
    body {
        background-color: #f7f9fc;
        font-family: "Arial", sans-serif;
    }
    .stButton>button {
        background-color: #1a73e8;
        color: white;
        justify-content: center;
        align-items: center;
        border-radius: 10px;
        padding: 10px;
        margin: 5px;
    }
    .stCheckbox {
        margin-top: 20px;
    }
    </style>
    """,
    unsafe_allow_html=True,
)

# Display welcome image
welcome_image_path = "bismillah.png"
if os.path.exists(welcome_image_path):
    st.image(welcome_image_path, use_container_width=True, caption="Bismillah hir Rehman Ar Raheem")
else:
    st.warning("Welcome image not found! Please add 'bismillah.png' in the script directory.")

st.title("Object Detection & Assistive Vision App for Visually Impaired People")
st.write("This application provides real-time object recognition and optional audio alerts.")

# Directory to store temp audio files
audio_temp_dir = "audio_temp_files"
if not os.path.exists(audio_temp_dir):
    os.makedirs(audio_temp_dir)

# Placeholder for video frames
stframe = st.empty()


# User controls
col1, col2 = st.columns(2)
with col1:
    start_detection = st.button("Start Detection")
with col2:
    stop_detection = st.button("Stop Detection")
audio_activation = st.checkbox("Enable Audio Alerts", value=False)

# Categories for audio alerts
alert_categories = {"person", "cat", "dog", "knife", "fire", "gun"}

# Dictionary to store the last alert timestamp for each object
last_alert_time = {}
alert_cooldown = timedelta(seconds=10)  # 10-second cooldown for alerts


def play_audio_alert(label, position):
    """Generate and play an audio alert."""
    phrases = [
        f"Be careful, there's a {label} on your {position}.",
        f"Watch out! {label} detected on your {position}.",
        f"Alert! A {label} is on your {position}.",
    ]
    alert_text = random.choice(phrases)

    temp_audio_path = os.path.join(audio_temp_dir, f"alert_{datetime.now().strftime('%Y%m%d_%H%M%S_%f')}.mp3")
    tts = gTTS(alert_text)
    tts.save(temp_audio_path)

    try:
        playsound(temp_audio_path)
        os.remove(temp_audio_path)  # Clean up after playing
    except Exception as e:
        print(f"Audio playback error: {e}")


def process_frame(frame, audio_mode):
    """Process a single video frame for object detection."""
    results = yolo(frame)
    result = results[0]

    detected_objects = {}
    for box in result.boxes:
        x1, y1, x2, y2 = map(int, box.xyxy[0])
        label = result.names[int(box.cls[0])]

        if audio_mode and label not in alert_categories:
            continue

        frame_center_x = frame.shape[1] // 2
        obj_center_x = (x1 + x2) // 2
        position = "left" if obj_center_x < frame_center_x else "right"

        detected_objects[label] = position

        cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 2)
        cv2.putText(
            frame,
            f"{label}",
            (x1, y1 - 10),
            cv2.FONT_HERSHEY_SIMPLEX,
            0.5,
            (0, 255, 0),
            2,
        )

    return detected_objects, frame


# Main logic
if start_detection:
    st.success("Object detection started.")
    try:
                detected_objects, processed_frame = process_frame(frame, audio_activation)

                frame_rgb = cv2.cvtColor(processed_frame, cv2.COLOR_BGR2RGB)
                stframe.image(frame_rgb, channels="RGB", use_column_width=True)

                if audio_activation:
                    current_time = datetime.now()
                    for label, position in detected_objects.items():
                        if (
                            label not in last_alert_time
                            or current_time - last_alert_time[label] > alert_cooldown
                        ):
                            play_audio_alert(label, position)
                            last_alert_time[label] = current_time

                time.sleep(0.1)

    except Exception as e:
        st.error(f"An error occurred: {e}")
    finally:
        if 'cap' in locals() and cap.isOpened():
            cap.release()
            cv2.destroyAllWindows()

elif stop_detection:
    st.warning("Object detection stopped.")