💡 Problem Formulation: In image processing, creating black and white images is a fundamental task that can be the starting point for various applications such as mask creation, background subtraction, or simply as placeholders. This article demonstrates five methods to create black (all pixels set to 0) and white (all pixels set to the maximum value, typically 255) images using OpenCV with Python. Our desired output is a solid black or white image of a given resolution, for example, 100×100 pixels.
Method 1: Using numpy.zeros()
In this method, we use the numpy library’s zeros function to create an array filled with zeros, which corresponds to a black image in OpenCV. The array’s shape should match the desired resolution and the number of color channels (1 for grayscale, 3 for RGB).
Here’s an example:
import numpy as np
import cv2
# Specify the resolution
width, height = 100, 100
# Create a black image
black_image = np.zeros((height, width, 3), np.uint8)
# Create a white image
white_image = np.ones((height, width, 3), np.uint8) * 255
# Display the images
cv2.imshow('Black Image', black_image)
cv2.imshow('White Image', white_image)
cv2.waitKey(0)
cv2.destroyAllWindows()Output:
This code snippet creates two windows displaying a solid black image and a solid white image, each of 100×100 pixels.
The numpy.zeros() function generates an array of a specified shape with all elements initialized to 0, creating a black image when displayed with OpenCV. We then use numpy.ones() multiplied by 255 to create a white image. Both arrays are cast to unsigned 8-bit integers (np.uint8).
Method 2: Using numpy.full()
Another approach involves the use of numpy’s full function that initializes an array with a specified shape and a fill value. This is efficient for creating a white image as you can directly specify the value 255 for each pixel.
Here’s an example:
import numpy as np
import cv2
# Specify the resolution
width, height = 100, 100
# Create a white image
white_image = np.full((height, width, 3), 255, dtype=np.uint8)
# Display the image
cv2.imshow('White Image', white_image)
cv2.waitKey(0)
cv2.destroyAllWindows()Output:
This code snippet creates a window displaying a solid white image of 100×100 pixels.
The numpy.full() function creates an array filled with the specified value (255 for white), and we specify the dtype as np.uint8 to match the data type OpenCV works with for images. This method is more direct compared to multiplying an array of ones and is more readable when creating a white image.
Method 3: Using cv2.rectangle()
This method uses OpenCV’s cv2.rectangle() function to draw a rectangle that spans the entire image, effectively filling it with color. By specifying a black or white color, you can fill the image accordingly. This is particularly useful when starting with an existing image or background.
Here’s an example:
import numpy as np
import cv2
# Specify the resolution
width, height = 100, 100
# Create a black image
black_image = np.zeros((height, width, 3), np.uint8)
cv2.rectangle(black_image, (0, 0), (width - 1, height - 1), (0, 0, 0), -1)
# Create a white image
white_image = np.zeros((height, width, 3), np.uint8)
cv2.rectangle(white_image, (0, 0), (width - 1, height - 1), (255, 255, 255), -1)
# Display the images
cv2.imshow('Black Image', black_image)
cv2.imshow('White Image', white_image)
cv2.waitKey(0)
cv2.destroyAllWindows()Output:
This code snippet generates two windows displaying a black image and white image respectively.
The cv2.rectangle() function is called with the coordinates of the image corners, color, and a fill value of -1 to draw a filled rectangle. This method is powerful when modifications need to be made to an existing image, as you can paint over it with a desired color.
Method 4: Using cv2.copyMakeBorder()
OpenCV’s cv2.copyMakeBorder() can also be used to create images by making a border around an existing image. If we start with a small kernel and create a large border, we effectively create a new image. This is a less common approach but useful for understanding OpenCV’s border-making capabilities.
Here’s an example:
import cv2
import numpy as np
# Small kernel
kernel = np.zeros((1, 1, 3), np.uint8)
# Specify the desired image size
width, height = 100, 100
# Create a black image
black_image = cv2.copyMakeBorder(kernel, 0, height-1, 0, width-1, cv2.BORDER_CONSTANT, value=(0, 0, 0))
# Create a white image
white_image = cv2.copyMakeBorder(kernel, 0, height-1, 0, width-1, cv2.BORDER_CONSTANT, value=(255, 255, 255))
# Display the images
cv2.imshow('Black Image', black_image)
cv2.imshow('White Image', white_image)
cv2.waitKey(0)
cv2.destroyAllWindows()Output:
This code snippet creates a black and a white image by adding borders to a 1×1 pixel core array.
The cv2.copyMakeBorder() function is used here creatively; by specifying a large border on a tiny 1×1 array and a constant value, we create a solid black or white image. This method is more of a clever hack than a direct approach, and it enriches the understanding of border controls within OpenCV.
Bonus One-Liner Method 5: Direct Array Initialization
For those who prefer brevity, directly initializing an array inline is the quickest way to create a black or white image in OpenCV. Here, we’ll use numpy to create a 100×100 black or white image in just one line.
Here’s an example:
import numpy as np
import cv2
# One-liner to create a black image
black_image = np.zeros((100, 100, 3), dtype=np.uint8)
# One-liner to create a white image
white_image = np.full((100, 100, 3), 255, dtype=np.uint8)
# Display the images
cv2.imshow('Black Image', black_image)
cv2.imshow('White Image', white_image)
cv2.waitKey(0)
cv2.destroyAllWindows()Output:
This code snippet succinctly creates and displays a black and white image of 100×100 pixels each.
These one-liners show the utility and simplicity of numpy when working with OpenCV. The np.zeros() and np.full() functions are powerful for creating images with uniform pixel values quickly and efficiently.
Summary/Discussion
- Method 1: Using numpy.zeros(). Strengths: Simple and intuitive way to create a black image. Weaknesses: Requires an additional step to create a white image.
- Method 2: Using numpy.full(). Strengths: Direct specification of fill value makes this perfect for creating white images. Weaknesses: Not as intuitive as numpy.zeros() for black images.
- Method 3: Using cv2.rectangle(). Strengths: Offers flexibility to modify existing images. Weaknesses: Slightly more verbose and less direct than methods using numpy.
- Method 4: Using cv2.copyMakeBorder(). Strengths: Demonstrates creative use of OpenCV functions. Weaknesses: Uncommon and somewhat convoluted method for this specific task.
- Bonus Method 5: Direct Array Initialization. Strengths: Extremely concise for both black and white image creation. Weaknesses: Does not explain the process, less educational for beginners.