Counting parking spots
Just like analytics of objects detected in aerial drone images is done just-in-time of query prompting and not for every drone image extracted from a drone video, unoccupied parking spots qualify as reasonable detections to be asked in a query. The corresponding analytics could take one of the following approaches:
Using color segmentation, contour detection and simple heuristics assuming that parking spots are regions in the parking lot not covered by cars.
import cv2
import numpy as np
def count_empty_parking_spots(image_path, debug=False):
image = cv2.imread(image_path)
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# Segment pavement (brightness, not car colors)
# Pavement is usually light gray under sunlight in aerial images
lower_pavement = np.array([0, 0, 180]) # min hue/sat, high V (brightness)
upper_pavement = np.array([180, 35, 255]) # low saturation to ignore cars
mask_pavement = cv2.inRange(hsv, lower_pavement, upper_pavement)
# Remove small noise
kernel = np.ones((7, 7), np.uint8)
mask_clean = cv2.morphologyEx(mask_pavement, cv2.MORPH_OPEN, kernel)
mask_clean = cv2.morphologyEx(mask_clean, cv2.MORPH_CLOSE, kernel)
# Adaptive thresholding as backup (if pavement color varies)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
adaptive = cv2.adaptiveThreshold(
gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 25, -14)
combined = cv2.bitwise_and(mask_clean, adaptive)
# Find contours (Connected bright regions interpreted as parking spots)
contours, _ = cv2.findContours(combined, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
empty_count = 0
display = image.copy()
# Filter contours: parking spots should be rectangular and large enough
for contour in contours:
area = cv2.contourArea(contour)
x, y, w, h = cv2.boundingRect(contour)
aspect = float(w) / float(h) if h > 0 else 0
# Tune these heuristics to ignore stray bright patches & small noise
if 400 < area < 2500 and 1.3 < aspect < 2.7 and w > 20 and h > 11:
empty_count += 1
cv2.rectangle(display, (x, y), (x+w, y+h), (0,255,0), 2)
print(f"Detected {empty_count} empty parking spots")
if debug:
cv2.imshow("Parking Detection", display)
cv2.imshow("Mask", mask_clean)
cv2.imshow("Adaptive", adaptive)
cv2.imshow("Combined", combined)
cv2.imwrite(image_path.replace(".jpg","")+"-detectedspots.jpg", display)
cv2.waitKey(0)
cv2.destroyAllWindows()
return empty_count
# Example usage for your two frames:
count1 = count_empty_parking_spots("parking1.jpg", debug=True)
count2 = count_empty_parking_spots("parking2.jpg", debug=True)
print(f"parking1.jpg: {count1} empty spots (should be about 2)")
print(f"parking2.jpg: {count2} empty spot (should be about 1)")
"""
Results: Detected 1 empty parking spots
Detected 1 empty parking spots
parking1.jpg: 1 empty spots (should be about 2)
parking2.jpg: 1 empty spot (should be about 1)
"""
Using YOLOv5 for speed, precision and adaptability to aerial views:
import torch
import cv2
import numpy as np
from matplotlib import pyplot as plt
# Load YOLOv5 model (you can use 'yolov5s', 'yolov5m', or a custom-trained model)
model = torch.hub.load('ultralytics/yolov5', 'yolov5s', pretrained=True)
# Load image
image_path = 'parking2.jpg'
img = cv2.imread(image_path)
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# Run inference
results = model(img_rgb)
# Parse results
detections = results.pandas().xyxy[0]
cars = detections[detections['name'] == 'car']
# Optional: If you have predefined parking spot coordinates
# Define parking spots manually or load from a JSON file
parking_spots = [
{'id': 1, 'bbox': [815, 184, 992, 387]},
{'id': 2, 'bbox': [230, 502, 273, 619]},
{'id': 3, 'bbox': [427, 8, 468, 691]},
# Add more spots...
]
# Check overlap between cars and parking spots
def is_occupied(spot_bbox, car_bboxes):
x1, y1, x2, y2 = spot_bbox
spot_area = (x2 - x1) * (y2 - y1)
for _, car in car_bboxes.iterrows():
cx1, cy1, cx2, cy2 = car[['xmin', 'ymin', 'xmax', 'ymax']]
ix1, iy1 = max(x1, cx1), max(y1, cy1)
ix2, iy2 = min(x2, cx2), min(y2, cy2)
iw, ih = max(0, ix2 - ix1), max(0, iy2 - iy1)
intersection = iw * ih
if intersection / spot_area > 0.3: # Threshold for overlap
return True
return False
# Count unoccupied spots
unoccupied_count = 0
for spot in parking_spots:
if not is_occupied(spot['bbox'], cars):
unoccupied_count += 1
print(f"Unoccupied parking spots: {unoccupied_count}")
The results of the second option appear as
YOLOv5 2025-7-21 Python-3.13.1 torch-2.7.1+cpu CPU
Fusing layers...
YOLOv5s summary: 213 layers, 7225885 parameters, 0 gradients, 16.4 GFLOPs
Adding AutoShape...
C:\Users\ravib/.cache\torch\hub\ultralytics_yolov5_master\models\common.py:906: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(autocast):
Unoccupied parking spots: 3
And the given input image is:
