How to Perform Adaptive Mean and Gaussian Thresholding of an Image using Python OpenCV?

OpenCV (Open Source Computer Vision) is an open-source library that provides tools for image and video analysis. In this tutorial, we will learn how to perform adaptive thresholding of an image using OpenCV in Python. Adaptive thresholding is a method used for image segmentation, where the threshold is computed for each pixel based on the values of the neighboring pixels. There are two types of adaptive thresholding: mean thresholding and Gaussian thresholding. In this tutorial, we will explore both of them.

Prerequisites

• Python 3
• OpenCV
• Numpy

First, we need to install the above-mentioned libraries. We can install them using the following pip command in the terminal:

pip install opencv-python numpy

Adaptive Mean Thresholding

Adaptive mean thresholding is a method where the threshold value is decided based on the mean value of the neighboring pixels. We specify a block size and a constant value, which is subtracted from the mean value of the block. If the pixel value is greater than the threshold, it is set to white, else it is set to black. The syntax for adaptive mean thresholding is as follows:

import cv2
import numpy as np

img = cv2.imread('image.jpg', 0)

# Adaptive Thresholding
thresh_mean = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 11, 2)

cv2.imshow('Adaptive Mean Thresholding', thresh_mean)
cv2.waitKey(0)
cv2.destroyAllWindows()

In the above code, we first read an image in grayscale. Then, we use the cv2.adaptiveThreshold() function to perform adaptive mean thresholding. The parameters used are explained below:

• img: The input image
• 255: The maximum value to be set for the threshold
• cv2.ADAPTIVE_THRESH_MEAN_C: The threshold method used, which is mean here
• cv2.THRESH_BINARY: The type of thresholding to be used
• 11: The size of the pixel neighborhood used to compute the threshold value
• 2: The constant value subtracted from the mean

We then display the output using the cv2.imshow() function.

Adaptive Gaussian Thresholding

Adaptive Gaussian thresholding is a method where the threshold value is decided based on the weighted average of the neighboring pixels using a Gaussian window. We specify a block size and a constant value, which is subtracted from the weighted average. If the pixel value is greater than the threshold, it is set to white, else it is set to black. The syntax for adaptive Gaussian thresholding is as follows:

import cv2
import numpy as np

img = cv2.imread('image.jpg', 0)

# Adaptive Thresholding
thresh_gaussian = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)

cv2.imshow('Adaptive Gaussian Thresholding', thresh_gaussian)
cv2.waitKey(0)
cv2.destroyAllWindows()

In the above code, we first read an image in grayscale. Then, we use the cv2.adaptiveThreshold() function to perform adaptive Gaussian thresholding. The parameters used are similar to those used in adaptive mean thresholding, except for the threshold method. Here, we use cv2.ADAPTIVE_THRESH_GAUSSIAN_C as the threshold method, which specifies the use of a Gaussian window.

Comparison between Mean and Gaussian Thresholding

To compare the performance of mean and Gaussian thresholding, we can use the following code:

import cv2
import numpy as np

img = cv2.imread('image.jpg', 0)

# Adaptive Mean Thresholding
thresh_mean = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 11, 2)

# Adaptive Gaussian Thresholding
thresh_gaussian = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)

cv2.imshow('Adaptive Mean Thresholding', thresh_mean)
cv2.imshow('Adaptive Gaussian Thresholding', thresh_gaussian)
cv2.waitKey(0)
cv2.destroyAllWindows()

In the above code, we first read the image in grayscale. Then, we perform both adaptive mean thresholding and adaptive Gaussian thresholding and display the outputs using the cv2.imshow() function.

Conclusion

In this tutorial, we learned how to perform adaptive thresholding of an image using OpenCV in Python using both meanand Gaussian thresholding. We saw how to implement them in Python using the cv2.adaptiveThreshold() function and the different parameters used for each method.

We also compared the performance of mean and Gaussian thresholding using a sample image. It is important to choose the appropriate method based on the type of image and the desired segmentation.

OpenCV provides a wide range of functions and tools for image and video analysis, making it a useful library for computer vision applications. With the knowledge gained from this tutorial, you can explore more advanced image processing techniques using OpenCV in Python.