Python | Bilateral Filtering (original) (raw)

Last Updated : 15 Jun, 2026

Bilateral filtering is an edge-preserving smoothing technique used in image processing to reduce noise while maintaining sharp boundaries in an image. It is widely used in computer vision tasks where preserving structural details is important.

• Reduces image noise without blurring important edges
• Uses both spatial and intensity information to preserve image structure

Working Principle

Bilateral filtering smooths an image by computing a weighted average of nearby pixels. The weights are based on two factors:

• **Spatial closeness: Nearby pixels have higher influence on the output
• **Intensity similarity: Pixels with similar brightness contribute more

This ensures that only similar pixels are blended, while edges (high intensity differences) are preserved. The filter behaviour is controlled by:

• \sigma_s (spatial parameter) that defines how far neighboring pixels are considered
• \sigma_r (range parameter) that controls how strongly intensity differences affect smoothing

If \sigma_r is large, the filter behaves similarly to a Gaussian blur.

Implementation

The following example demonstrates the use of OpenCV's bilateral filter for image smoothing. The filter reduces noise while maintaining important edges, and the result is compared with the original image.

**Input Image:

Python `

import cv2

img = cv2.imread('taj.jpg')

bilateral = cv2.bilateralFilter(img, 15, 75, 75)

cv2.imwrite('taj_bilateral.jpg', bilateral)

`

**Output:

Advantages

• Preserves edges while removing noise effectively
• Works well for natural images where boundaries are important
• Produces visually sharper results compared to Gaussian blur

Limitations

• Computationally slower than basic smoothing filters
• Performance depends heavily on parameter selection
• Not ideal for real-time processing on large images without optimization