What is the Canny edge detection technique in OpenCV

Recipe Objective: What is the Canny edge detection technique in OpenCV?

This recipe will help you understand the Canny edge detection technique in OpenCV

Step 1: Import the libraries and read the image

Let us first import the necessary libraries and read the image. The image that we are using here is the one shown below. Let us read the image in grayscale this time as Canny works only on grayscale images.

import numpy as np
import cv2
image = cv2.imread('edgeflower.jpg',0)


Step 2: Understanding the theory of the Canny edge detection algorithm

The Canny edge detection algorithm was first invented by John F. Canny and hence the name Canny algorithm. This algorithm involves a series of steps which we will discuss in this step.

1. The first step is to reduce the noise and smoothen the image using a 5 x 5 Gaussian filter.
2. The second step is to find the intensity gradient of the image using a Sobel kernel.
3. The third step is called Non - maximum suppression. In this step, each pixel is scanned individually, and non-edge pixels are removed. Each point is checked if it is the local maximum of its neighborhood. If yes, then it is detected as an edge, else omitted.
4. The last step is Hysteresis Thresholding. This is the previous step which decides if the pixel is an edge pixel or not. This takes a maxVal and a minVal. The pixels edges with an intensity gradient more than the maxVal are considered as a definite edge. Similarly, the ones below the minVal are regarded as a definite non-edge pixel. The pixels between the maxVal and minVal are categorized relative to the definite edges.

Step 3: Implementation of Canny edge detection

All the steps mentioned above are condensed into one function called cv2.Canny() takes the following arguments.

• image: The input image in grayscale
• threshold1: The value for minVal
• threshold2: The value for maxVal

Here's how we can use it for our image

canny_edges = cv2.Canny(image,100,190)

Step 4: Display the output

Let us display the detected output using matplotlib

plt.figure(figsize=(13,5))
plt.subplot(1,2,1)
plt.imshow(image,cmap = 'gray')
plt.title('Original Image')
plt.xticks([])
plt.yticks([])

plt.subplot(1,2,2)
plt.imshow(canny_edges,cmap = 'gray')
plt.title('Edges detected using Canny algorithm')
plt.xticks([]), plt.yticks([])

plt.show()
plt.tight_layout()


Output: