OpenCV

OpenCV Histograms and Color Segmentation

# Read the image.
img = cv2.imread('Apollo-8-Launch.jpg', 0)

# Use calcHist() in OpenCV.
hist = cv2.calcHist(images = [img], channels = [0], mask = None, histSize = [256], ranges = [0,255])

# Flatten the image data.
img_flatten = img.ravel()

# Display the image and histograms.
plt.figure(figsize = [18, 4])
plt.subplot(131); plt.imshow(img); plt.title('Original Image')
plt.subplot(132); plt.plot(hist); plt.xlim([0, 256]); plt.title('cv2.calcHist()')
plt.subplot(133); plt.hist(img_flatten,256,[0,256]); plt.xlim([0, 256]); plt.title('np.ravel(), plt.hist()')

# Read the color image.
img = cv2.imread('parrot.jpg', cv2.IMREAD_COLOR)

# Convert to HSV.
img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)

# Perform histogram equalization only on the V channel, for value intensity.
img_hsv[:,:,2] = cv2.equalizeHist(img_hsv[:, :, 2])

# Convert back to BGR format.
img_eq = cv2.cvtColor(img_hsv, cv2.COLOR_HSV2BGR)

# Display the images.
plt.figure(figsize = (18, 6))
plt.subplot(121); plt.imshow(img[:, :, ::-1]); plt.title('Original Color Image')
plt.subplot(122); plt.imshow(img_eq[:, :, ::-1]); plt.title('Equalized Image')

# Read the image in a 3 channel color format.
img = cv2.imread('opencv_logo.png', cv2.IMREAD_COLOR)
img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)

# Set range for red color.
r_lb = np.array([165, 50, 50], np.uint8)
r_ub = np.array([180, 255, 255], np.uint8)

# Set range for green color.
g_lb = np.array([35, 50, 50], np.uint8)
g_ub = np.array([80, 255, 255], np.uint8)

# Set range for blue color.
b_lb = np.array([95, 50, 50], np.uint8)
b_ub = np.array([125, 255, 255], np.uint8)

# Define each color mask.
r_mask = cv2.inRange(img_hsv, r_lb, r_ub)
g_mask = cv2.inRange(img_hsv, g_lb, g_ub)
b_mask = cv2.inRange(img_hsv, b_lb, b_ub)

# Display each color mask.
plt.figure(figsize = (18, 4))
plt.subplot(131); plt.imshow(r_mask); plt.title('Red Mask')
plt.subplot(132); plt.imshow(g_mask); plt.title('Green Mask')
plt.subplot(133); plt.imshow(b_mask); plt.title('Blue Mask');

# Segment the colors.
r_seg = cv2.bitwise_and(img, img, mask = r_mask)
g_seg = cv2.bitwise_and(img, img, mask = g_mask)
b_seg = cv2.bitwise_and(img, img, mask = b_mask)

# Display the segmented colors.
plt.figure(figsize = (18, 4))
plt.subplot(131); plt.imshow(r_seg[:, :, ::-1]); plt.title('Red Color Segmented')
plt.subplot(132); plt.imshow(g_seg[:, :, ::-1]); plt.title('Green Color Segmented')
plt.subplot(133); plt.imshow(b_seg[:, :, ::-1]); plt.title('Blue Color Segmented');

OpenCV Histograms and Color Segmentation

References