Gaussian Blur

Updated 02 Oct 2026

•
What it is
- •
  smoohs an image using weights sampled from a gaussian function
•
Formula
- •
  before discrete normalization, $h(x,y)\propto e^{-(x^2+y^2)/(2\sigma^2)}$ .
•
Parameters
- •
  $\sigma$ controls the spread of the weights
- •
  kernel size controls how much of the Gaussian is retained.
- •
  normalize the discrete weights to sum to $1$ .
•
Interpretation
- •
  reduces rapid spatial changes, including some noise and fine detail
- •
  does not distinguish unwanted noise from useful high-frequency structure
•
Project
- •
  the low-pass branch uses GaussianBlur(image,(5,5),0) asking OpenCV to dervie sigma from the kernel size
•
Related
- •
  Mean Filter
- •
  Frequency-Domain Filtering

Linked references 4