car_counter.py

import cv2
from ultralytics import YOLO

# 1. Initialize the YOLO model (Pre-trained on 80 common object classes)
# 'yolov8n.pt' is the nano model - incredibly fast and ideal for CPUs/laptops
model = YOLO("yolov8n.pt")

# Define target COCO dataset class IDs for vehicles we want to count
# Class 2 = car, Class 3 = motorcycle, Class 5 = bus, Class 7 = truck
VEHICLE_CLASSES = [2, 3, 5, 7]

# 2. Choose your input source:
# For a video file: put the video in your project folder and replace with "traffic.mp4"
# For live webcam/traffic camera feed: use 0
source ="video.mp4" 

cap = cv2.VideoCapture(source)

print("\n🚀 Vehicle Counter Active! Press 'q' to quit.")

while cap.isOpened():
    success, frame = cap.read()
    if not success:
        print("End of video stream or camera unavailable.")
        break

    # 3. Pass the frame to YOLO (stream=True optimizes memory for loops)
    # verbose=False keeps our terminal clean from clutter
    results = model(frame, stream=True, verbose=False)
    
    car_count = 0

    # 4. Process the detection results
    for result in results:
        boxes = result.boxes  # Box coordinates for detected items
        
        for box in boxes:
            # Extract the class ID integer
            cls_id = int(box.cls[0])
            
            # If the detected object is a car, truck, bus, or motorcycle
            if cls_id in VEHICLE_CLASSES:
                car_count += 1
                
                # Get bounding box coordinates: Top-Left (x1, y1) and Bottom-Right (x2, y2)
                x1, y1, x2, y2 = map(int, box.xyxy[0])
                
                # Extract the confidence score (e.g., 0.85 means 85% sure it's a car)
                confidence = float(box.conf[0])
                
                # Get the human-readable class name text (e.g., "car")
                class_name = model.names[cls_id]
                
                # Draw a bounding box rectangle around the vehicle (Green color)
                cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 2)
                
                # Draw a text label stating the object type and confidence score
                label = f"{class_name.upper()} {confidence:.2f}"
                cv2.putText(frame, label, (x1, max(0, y1 - 10)), 
                            cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)

    # 5. Display the overall tally count on the Screen (HUD Overlay)
    hud_text = f"Total Vehicles Spotted: {car_count}"
    # Draw a shadow outline first for high contrast readability
    cv2.putText(frame, hud_text, (22, 52), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 0), 4, cv2.LINE_AA)
    cv2.putText(frame, hud_text, (20, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 0), 2, cv2.LINE_AA)

    # Show the output frame window
    cv2.imshow("YOLO Real-Time Vehicle Analytics", frame)

    # Exit window gracefully on pressing 'q'
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

# Cleanup
cap.release()
cv2.destroyAllWindows()