Adding a Transparent Watermark to Images with OpenCV

Overview

Adding a watermark to images is a common task in digital image processing, useful for branding, copyright protection, or simply marking your work. In this tutorial, you'll learn how to overlay a transparent watermark onto another image using OpenCV and Python, with step-by-step explanations and code.

Prerequisites

You'll need the following Python libraries:

• opencv-python: For image processing operations
• matplotlib: For displaying images
• numpy: For numerical operations

Install them with:

pip install opencv-python matplotlib numpy

Basic Setup

Let's start by importing the necessary libraries:

import sys
import numpy as np
np.set_printoptions(threshold=sys.maxsize)
import cv2
from matplotlib import pyplot as plt

Step 1: Reading and Preparing Images

Loading Images

We'll use two images: a main image and a watermark image.

# Read Images in BGR Mode
dogImg = cv2.imread('Assets/dog_backpack.jpg')
watermarkImg = cv2.imread('Assets/watermark_no_copy.png')

Converting Color Spaces

OpenCV loads images in BGR format. Convert them to RGB for correct display:

dogImg = cv2.cvtColor(dogImg, cv2.COLOR_BGR2RGB)
watermarkImg = cv2.cvtColor(watermarkImg, cv2.COLOR_BGR2RGB)

Resizing the Watermark

Resize the watermark to fit the target area:

watermarkImg = cv2.resize(watermarkImg, (600, 600))
plt.imshow(watermarkImg)
plt.show()

Step 2: Defining the Region of Interest (ROI)

To overlay the watermark, we need to select a region in the main image where the watermark will be placed.

# Calculate offsets for placement
x_offset = 934 - 600
y_offset = 1401 - 600

# Get watermark dimensions
rows, cols, channels = watermarkImg.shape

# Define ROI in the main image
roi = dogImg[y_offset:1401, x_offset:984]

Step 3: Creating a Mask for the Watermark

Convert Watermark to Grayscale

We'll use a mask to separate the watermark from its background.

watermarkGray = cv2.cvtColor(watermarkImg, cv2.COLOR_RGB2GRAY)
plt.imshow(watermarkGray, cmap='gray')
plt.show()

Create the Mask

Remove the white background of the watermark image, bitwise_not: meaning it converts all 0s to 1s and all 1s to 0s.

Invert the grayscale watermark to create a mask:

mask = cv2.bitwise_not(watermarkGray)
plt.imshow(mask, cmap='gray')
plt.show()

Step 4: Preparing Background and Foreground

Create a White Background

whiteBackground = np.full(watermarkImg.shape, 255, dtype=np.uint8)

Apply the Mask

Use bitwise operations to isolate the watermark and its background:

# White background with mask
bk = cv2.bitwise_or(whiteBackground, whiteBackground, mask=mask)
plt.imshow(bk)
plt.show()

# Watermark foreground with mask
fg = cv2.bitwise_or(watermarkImg, watermarkImg, mask=mask)
plt.imshow(fg)
plt.show()

Step 5: Combining the Watermark with the ROI

Overlay the watermark onto the selected region:

final_roi = cv2.bitwise_or(roi, fg)
plt.imshow(final_roi)
plt.show()

Step 6: Placing the Watermarked ROI Back

Paste the watermarked region back into the original image:

large_img = dogImg
small_img = final_roi

large_img[y_offset:y_offset+small_img.shape[0], x_offset:x_offset+small_img.shape[1]] = small_img

plt.imshow(large_img)
plt.show()

Understanding the Process

• ROI (Region of Interest): The area in the main image where the watermark will be placed.
• Masking: Used to separate the watermark from its background, allowing for transparent overlays.
• Bitwise Operations: Essential for combining images and masks in OpenCV.