To work with OpenCV from python, you need to install it first:
pip install opencv-python
After we import cv2 we can directly work with images like so:
import cv2 img = cv2.imread("doc_brown.png")
For showing the image, it is recommended to use matplotlib
import matplotlib.pyplot as plt img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) plt.imshow(img) plt.show(
OpenCV stores images internally in the BGR format – blue – green – red so we have to convert to RGB before displaying them.
We have now an image object which can already tell us more about the image itself:
print(img.shape) (205, 236, 3)
The tuple shows us the number of (rows, columns, channels)
import numpy as np brightness = np.zeros(img.shape, dtype="uint8") + 30 img = cv2.add(img, brightness)
Manipulating brightness works like this: you need a numpy array with the size of the image, add your brightness and add the brightness array to the original image.
Adding noise works the same way, but instead of adding a fixed value to every pixel you add normal distributed values.
noise = np.random.randint(0, 100, size=img.shape, dtype="uint8") img = cv2.add(img, noise)
Smoothing is reducing the noise of an image by adding a Gaussian blur:
img = cv2.GaussianBlur(img, ksize=(31, 31), sigmaX=5)
Last but not least we can convert an image to a gray scale. Beware that for showing / storing the image you need to add the flag cmap=”gray”
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) plt.imshow(img, cmap='gray')
Have fun fiddling around with OpenCV!