The Image Processing Series Part 3: Drawing with OpenCV

Filed Under: Python
Opencv Part3 Min

In this article, we’ll talk about Drawing with OpenCV. If you haven’t done so already, be sure to read the previous article on setting pixels with Python OpenCV.

A large part of working with Image Processing has to do with identifying objects and contours in a provided image.

This can be done by marking these shapes during the processing.

But, how do we do so?

The same question that plagued the scientists working on Image Processing led to the development of drawing in OpenCV.

We’ll be using a few of these functions to begin working with drawing in OpenCV in this article.

Note that this is the third article of the Image Processing Series, and it might a good idea to look into the introductory session before proceeding through this article!

Drawing with Python OpenCV

Let’s start drawing with the OpenCV module.

Setting it all up

In order to get started, we’ll first need to import the modules that we’ll be using for drawing in OpenCV.

This, of course, consists of the cv2 module, but, another module we’ll be using is the numpy module.

In case you don’t have numpy already installed in your system, here’s the command to install it via pip:

pip install numpy

We’ll then proceed to utilize these modules in the code by importing them into the script:

# importing the numpy module to create a canvas of 0 value pixels.
import numpy as np

# importing the OpenCV module
import cv2

In order to draw anything, we’ll first need to have something to draw upon, so, let’s initialize a canvas to do so.

We’ll be doing this by creating a canvas using the Numpy module and make one that has 300×300 pixels with a value of each pixel set to 0.

This means that we’ll be presented with a 900 pixel black canvas, as the value for 0 is black in Images.

# setting up a black canvas of 900 pixels
canvas = np.zeros((300, 300, 3), dtype="uint8")

We’ll also set up a color to a variable that allows us to use it when we draw objects in the canvas.

# setting the color green to a variable.
green = (0, 255, 0)

Now that we’re all set up, let’s get to drawing items on the canvas!

Let’s draw lines

We’ll now draw two lines across both the diagonals using the line() function.

To show the difference between thickness, we’ll be changing up the color for the other diagonal.

# To draw a green line across the canvas whose size you have to adjust through the editor?
cv2.line(canvas, (0, 0), (300, 300), green)
cv2.imshow("Canvas", canvas)

# Draws a three-inch red line across the other diagonal
red = (0, 0, 255)
cv2.line(canvas, (300, 0), (0, 300), red, 3)
cv2.imshow("Canvas", canvas)

The output that you should receive should look sometime like this,

Lines on the canvas

Working with Boxes

Proceeding forwards, we can use the rectangle() function from the cv2 module to draw rectangles in the canvas.

# While this command draws a rectangle given the right dimensions
cv2.rectangle(canvas, (10, 10), (60, 60), green)
cv2.imshow("Canvas", canvas)

# Draws a rectangle when specified the co-ordinates, with a thickness for the edges
cv2.rectangle(canvas, (20, 20), (70, 70), red, 5)
cv2.imshow("Canvas Red", canvas)
# The duration for which the canvas stays by the defined value

blue = (255, 0, 0)
# -1 thickness causes the entire box to get filled.
cv2.rectangle(canvas, (200, 50), (225, 125), blue, -1)
cv2.imshow("Canvas Blue", canvas)

The output that you should be receiving upon doing so, should look something like this:

Image 1
Drawing using Rectangles on the Canvas

Circles and Randomization

Mixing things up, we’ll now proceed to working with circles. Here’s two examples of working with circles.

Concentric Circles

Here, we are centering the point of origin for all the circles, and drawing them in increments of the radius.

# centers the canvas to the center and forms rings based on the increment of the radius
canvas = np.zeros((300, 300, 3), dtype="uint8")
(centerX, centerY) = (int(canvas.shape[1] / 2), int(canvas.shape[0] / 2))
white = (255, 255, 255)
for r in range(0, 175, 25):, (centerX, centerY), r, white)
cv2.imshow("Canvas White", canvas)

The output that you receive, should look something like this,

Image 2
Concentric Circles on the canvas

Random Circles

Here, we are using the numpy module to randomize the size of the circles and the points at which they will be placed in the canvas.

We’ll also place a waitkey() such that we can view each circle individually before proceeding to the next one.

# Drawing circles on random with varying radii,
# throughout the canvas
for i in range(0, 25):
    radius = np.random.randint(5, high=200)
    color = np.random.randint(0, high=256, size=(3,)).tolist()
    pt = np.random.randint(0, high=300, size=(2,)), tuple(pt), radius, color, -1)

    cv2.imshow("Canvas", canvas)

Working with the circles, the output that you receive will be random, and as such, it’s better to test it out on your own 馃槈

Altering existing images

We can alter existing images, by loading them into a variable using the cv2.imread() function.

# reading the image into a variable using the imread() function
image = cv2.imread(
# drawing three circles and a rectangle on the image, (168, 188), 90, (0, 0, 255), -2), (150, 164), 10, (0, 0, 255), -1), (192, 174), 10, (0, 0, 255), -1)
cv2.rectangle(image, (134, 200), (186, 218), (0, 0, 255), -1)

# showing the output of the process
cv2.imshow("Output", image)

I’ve input an image which is present in the assets of the repository, but you might wish to use your own image.

Image 3
The output image which was modified from the assets.

This can be done by modifying the image location!


We’ve learnt how to perform drawing with OpenCV. The use cases for working with drawing lines in OpenCV are definitely not useless.

Throughout the course of working with Image Processing, we can draw contours on images, draw boxes while identifying objects, and use it for live tracking objects in videos.

The implementations of drawing in OpenCV allows for us to work with Machine Learning models in the future as well.

Want to look at the future where we work with Image Processing on a larger scale?

Look into OpenCV and Facial Recognition, and a Java implementation of Android and CameraX OpenCV!


Generic selectors
Exact matches only
Search in title
Search in content