Object Counter

Count objects in images using contour detection.

View Source Code

Overview

Automatically count objects like coins, pills, or items on a conveyor belt using image segmentation and contour analysis.

Key Techniques:

  • Adaptive thresholding
  • Morphological operations
  • Contour detection
  • Connected component analysis

How It Works 

Image → Grayscale → Threshold → Morphology → Find Contours → Count
   ↓        ↓           ↓           ↓             ↓           ↓
[RGB]   [Gray]     [Binary]    [Clean]      [Objects]    [N=15]

Key OpenCV Functions 

# Convert and threshold
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
_, binary = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)

# Clean up with morphology
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))
cleaned = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel)
cleaned = cv2.morphologyEx(cleaned, cv2.MORPH_CLOSE, kernel)

# Find contours
contours, _ = cv2.findContours(cleaned, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

# Filter by size and count
min_area = 100
objects = [c for c in contours if cv2.contourArea(c) > min_area]
count = len(objects)

Connected Components Alternative 

# Label connected regions
num_labels, labels, stats, centroids = cv2.connectedComponentsWithStats(binary)

# stats contains: x, y, width, height, area for each component
for i in range(1, num_labels):  # Skip background (0)
    x, y, w, h, area = stats[i]
    if area > min_area:
        cv2.rectangle(image, (x, y), (x+w, y+h), (0, 255, 0), 2)

Watershed for Touching Objects

When objects touch, use watershed segmentation:

# Distance transform
dist = cv2.distanceTransform(binary, cv2.DIST_L2, 5)
_, sure_fg = cv2.threshold(dist, 0.5 * dist.max(), 255, 0)

# Find markers
_, markers = cv2.connectedComponents(sure_fg.astype(np.uint8))

# Apply watershed
markers = cv2.watershed(image, markers)

Controls

Key Action
+/- Adjust threshold
m Toggle morphology
w Use watershed
s Save result
q Quit

Running the Application 

python curriculum/applications/05_object_counter.py

Official Documentation