Performing Computer Vision Tasks with OpenCV and Python

Computer vision has become an increasingly important field, with applications ranging from self-driving cars to facial recognition. One of the most popular tools for performing computer vision tasks is the OpenCV library, which provides a wide range of functions for image and video processing.

In this article, we‘ll explore how to use OpenCV with Python to perform some fundamental computer vision operations. We‘ll read in an image, examine its properties, manipulate the pixel values, convert it to grayscale, and more. By the end, you‘ll have a solid foundation for tackling your own computer vision projects using these versatile tools.

Representing Images as NumPy Arrays

When you use OpenCV to read an image into memory, it represents that image as a NumPy array. NumPy is a powerful Python library for numerical computing that allows you to efficiently store and manipulate large, multi-dimensional arrays.

In the case of images, the NumPy array will have a shape of (height, width, channels), where height is the number of pixel rows, width is the number of pixel columns, and channels is the number of color channels (e.g. 3 for RGB images or 1 for grayscale).

Each pixel is represented by one or more intensity values in the array, depending on the color space and data type. Common data types are unsigned 8-bit integers with a range of 0-255.

Reading an Image with OpenCV

Let‘s see how to use OpenCV to read in an image file. We‘ll use the OpenCV logo as an example:

import cv2

# Read image from disk
img = cv2.imread(‘opencv-logo.png‘)

# Print image dimensions
print(img.shape)

Output:

(739, 600, 3)  

The cv2.imread() function reads the image file specified by the path and returns it as a NumPy array. The .shape property reveals that this is a 739×600 pixel RGB color image.

Accessing and Manipulating Pixels

Now that we have the image in a NumPy array format, we can easily access and manipulate the pixel values using standard NumPy indexing and slicing:

# Get the pixel value at row 100, column 200
pixel = img[100, 200] 
print(pixel)

# Set the pixel to black
img[100, 200] = [0, 0, 0]

Output:

[255 255 255]

Here we access an individual pixel value using array indexing and print out its RGB values. We then modify that pixel by assigning it new RGB values.

We can also access larger chunks of the image at once using slicing:

# Extract a 100x100 pixel sub-region starting from row 200, column 100
sub_img = img[200:300, 100:200]

# Display the sub-region in a new window 
cv2.imshow(‘Sub-region‘, sub_img)
cv2.waitKey(0)

This extracts a 100×100 pixel rectangle from the original image and displays it in a new window. The cv2.waitKey(0) line causes the program to pause until a key is pressed, giving us a chance to view the image.

Converting to Grayscale

A common operation in computer vision is converting images to grayscale. This simplifies the image data from three color channels down to one intensity channel. OpenCV makes this easy with the cv2.cvtColor() function:

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

cv2.imshow(‘Grayscale‘, gray) 
cv2.waitKey(0)

The grayscale version of the image is created by passing the original image and the cv2.COLOR_BGR2GRAY flag to indicate the desired conversion. We can then display it in a new window as before.

Thresholding

Another useful technique is thresholding, which converts a grayscale image into a binary image by setting pixel values above a certain threshold to white and values below to black. This can help isolate objects of interest:

_, thresh = cv2.threshold(gray, 128, 255, cv2.THRESH_BINARY)

cv2.imshow(‘Thresholded‘, thresh)
cv2.waitKey(0)  

The first argument to cv2.threshold is the input grayscale image. The second is the threshold value. The third is the maximum value to use. And the last specifies the thresholding method, in this case, binary thresholding.

Exploring Further

We‘ve only scratched the surface of what‘s possible with OpenCV. It provides functions for tasks like:

• Edge detection
• Object detection and tracking
• Feature matching
• Face and eye detection
• Camera calibration
• Optical flow
• Much more

OpenCV combined with Python‘s ease of use and extensive libraries make it a great platform for rapidly prototyping and deploying computer vision applications. The online OpenCV documentation is the best place to start learning more about its extensive capabilities.

Conclusion

In this article, we‘ve seen how OpenCV represents images as NumPy arrays and allows you to easily read, display, and manipulate image data using Python. We demonstrated some basic operations like pixel access, slicing, colorspace conversion and thresholding.

Hopefully this taste of OpenCV has whetted your appetite to explore further. With its wide range of functions and Python‘s simplicity, you have a powerful toolkit for solving real-world computer vision problems. So dive in, experiment, and see what you can create! The possibilities are endless.