Simple Thresholding using OpenCV

Thresholding is a foundational technique in computer vision and image processing used to segment objects from the background. It works by comparing each pixel value of a grayscale image against a specified threshold value. Based on this comparison, pixels are assigned new values, usually 0 (black) or 255 (white).In OpenCV with Python, the function cv2.threshold is used for thresholding.

In thresholding, for every pixel at position (x,y) with an intensity value f(x,y):

If f(x,y) < T, set the pixel to 0 (black).
If f(x,y) ≥ T, set it to the maximum value (typically 255, white).
Here, T is the threshold value, and the process is usually performed on a grayscale version of the image

Technique	Description
cv2.THRESH_BINARY	Above threshold -> 255, below -> 0.
cv2.THRESH_BINARY_INV	Above threshold -> 0, below -> 255 (inverse of binary).
cv2.THRESH_TRUNC	Above threshold -> set to threshold, below -> unchanged.
cv2.THRESH_TOZERO	Below threshold -> 0, above -> unchanged.
cv2.THRESH_TOZERO_INV	Above threshold -> 0, below -> unchanged (inverse of TOZERO).

Step-by-Step Implementation

Let's implement the various types of simple thresholding techniques,

Step 1: Import libraries and Image Preparation

Sample image can be downloaded from here.

Let's import the required libraries and load our image on which we will perform the operations,

cv2: Handles image reading, processing, and applies thresholding techniques.
numpy: Supports efficient array operations, enabling fast image data handling.
matplotlib.pyplot: Displays images and results in Colab notebooks.

Python

import cv2
import numpy as np
import matplotlib.pyplot as plt

image = cv2.imread('input.png')
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

Step 2: Helper Function

Define the helper function which helps in displaying the images,

Python

def show_image(img, title):
    plt.imshow(img, cmap='gray')
    plt.title(title)
    plt.axis('off')
    plt.show()

Step 3: Display the Original Image

Python

show_image(gray_image, 'Original Grayscale Image')

Output: