OpenCV in Python

1 minute read

OpenCV is great, but it is not so great with the documentation for its Python binding.

Certainly, there are resourses and tutorials, but overall, Python is at most a second-class citizen in the OpenCV world due to lack of many API supports (e.g. missing many algorithm interface here. The upside is: Python may also not need some of them either (e.g. the machine learning part can resort to sklearn). The strategy of OpenCV in Python then becomes “identify useful things or find alternatives” until satisfying the need.

This article aims to fit some of such gaps for newbies.

Error Control

It seems not stated explicitly, but the Python binding has an exception class called cv2.error. The class is not following the overall OpenCV Python binding naming convention so it may be subject to change in the future.

Display in Jupyter Notebook


One can use the imshow function from matplotlib.pyplot:

import matplotlib.pyplot as plt
import cv2

img_bgr = cv2.imread('sample.jpg')  # OpenCV default: BGR
img_rgb = cv2.cvtColor(cv2.COLOR_BGR2RGB)

Note that OpenCV default is BGR for some history reason, whereas plt.imshow is in the RGB order. cv2.imshow, on the other hand, is consistent with cv2.imread in the BGR order.


Use ipython magic clear_output(wait=True) to update frame

from IPython.display import clear_output
import cv2
import numpy as np
import sys

vid = cv2.VideoCapture(0)

        # Capture frame-by-frame
        ret, frame =
        if not ret:
            raise cv2.error("Camera failed to capture")

        # Convert the image from OpenCV BGR format to matplotlib RGB format
        # to display the image
        frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        plt.axis('off')  # Turn off the axis

        # Display the frame until new frame is available
except KeyboardInterrupt:
    sys.stdout.write("Keyboard Interrupt\n")
except cv2.error as e:
    # Message to be displayed after releasing the device
    sys.stdout.write("Release Video Resource\n")


Leave a Comment