# Initialize components

detector = ObjectDetector(model_path="yolov8n.pt", frame_width=640, frame_height=320)

analyzer = SituationAnalyzer()

audio = AudioFeedback()

voice_control = VoiceController()

# Initialize video capture (0 is usually the built-in webcam)

cap = cv2.VideoCapture(1)

if not cap.isOpened():

print("Error: Could not open video capture device")

return

# Set camera resolution

cap.set(cv2.CAP_PROP_FRAME_WIDTH, 640)

cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 320)

print("Waiting for activation phrase: 'Go'...")

print("Press 'q' to quit")

last_analysis_time = 0

analysis_interval = 1.0 # Analysis every 1 second

last_detection_results = None

running = True

try:

while running:

# Capture frame

ret, frame = cap.read()

if not ret:

print("Error: Could not read frame")

continuepy

# Check if system is activated

if voice_control.is_system_active():

current_time = time.time()

# Only process frame at the specified interval

if current_time - last_analysis_time >= analysis_interval:

# Process frame with YOLO

detection_results = detector.process_frame(frame)

# Analyze situation only if detections have changed

if _has_detections_changed(last_detection_results, detection_results):

# Analyze situation

analysis = analyzer.analyze_situation(detection_results)

# Only provide audio feedback if the situation has changed

if analysis.get("changed", True):

audio.provide_feedback(analysis)

last_detection_results = detection_results

last_analysis_time = current_time

# Get audio feedback stats

stats = audio.get_stats()

# Add visual status indicators

status_color = (0, 255, 0) if analysis.get("safe_to_proceed", True) else (0, 0, 255)

cv2.putText(frame, f"ACTIVE - {analysis.get('guidance', '')}",

(10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.7, status_color, 2)

# Display audio feedback stats

delay_color = (0, 255, 0) if stats['last_message_delay'] < 0.5 else (0, 165, 255)

cv2.putText(frame, f"Audio Delay: {stats['last_message_delay']:.2f}s",

(10, frame.shape[0] - 60), cv2.FONT_HERSHEY_SIMPLEX, 0.6, delay_color, 2)

cv2.putText(frame, f"Queue Size: {stats['queue_size']}",

(10, frame.shape[0] - 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2)

# Always draw detection boxes on frame

if last_detection_results:

frame = detector.draw_detections(frame, last_detection_results)

else:

# Draw waiting message on frame

cv2.putText(frame, "Waiting for activation: 'Go'",

(10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)

# Show processed frame

cv2.imshow('Wheelchair Navigation Assistant', frame)

# Check for quit command

key = cv2.waitKey(1) & 0xFF

if key == ord('q'):

running = False

print("\nQuitting application...")

except KeyboardInterrupt:

print("\nStopping application...")

finally:

# Cleanup

print("Cleaning up resources...")

cap.release()

cv2.destroyAllWindows()

audio.cleanup()

"""Compare detection results to determine if there's a significant change."""

if last_results is None:

return True

last_detections = last_results.get("detections", [])

current_detections = current_results.get("detections", [])

# Compare number of detections

if len(last_detections) != len(current_detections):

return True

# Create simplified representations for comparison

def simplify_detection(det):

return (

det["object"],

det["distance"],

det["position"]

)

last_simplified = {simplify_detection(det) for det in last_detections}

current_simplified = {simplify_detection(det) for det in current_detections}