CoreSource checked

Canny Edge Detection

Finds thin edges by smoothing, measuring gradients, suppressing nonmaxima, and linking thresholds.

Explore relationships Open constellationBrowse atlas

Quick Facts

Compact metadata for implementation and discovery.

Category: Image Processing
Difficulty: Intermediate
Family: Edge Detection
Tier: Core
Status: Verified
Review: Source Checked
Sources: 2
Formulas: 1
Relationships: 7
Code example: Ready
Visualization: Ready
Aliases: Canny Edge Detector

Overview

Canny edge detection reduces noise with smoothing, computes gradient magnitude and direction, thins responses with non-maximum suppression, and uses hysteresis thresholds to keep connected strong edges.

Core Concepts

pixel operationgradient responsefilter kerneledge detectionpixel neighborhoodlocal operationimage transform

Quick facts

Complexity

Runtime and memory behavior at a glance.

Best

O(w h)

Average

O(w h)

Worst

O(w h)

Space

O(w h)

When to Use

Use when a clean, thin edge map is needed from grayscale intensity structure.
Use when threshold and smoothing parameters can be tuned for the image scale.

When Not to Use

Avoid when texture or noise dominates meaningful edges.
Avoid when downstream code needs learned semantic boundaries rather than intensity edges.

What It Does

Smooths the image before gradient measurement
Computes gradient magnitude and direction
Thins and links edge responses with hysteresis

Use Cases

Classical edge maps for vision pipelines
Shape extraction before contour analysis
Teaching multi-stage image-processing pipelines

Inputs

Grayscale image
Low and high thresholds

Outputs

Binary edge map

Key Formulas (1)

Gradient magnitude

magnitude = √gx² + gy²

magnitude = sqrt(gx^2 + gy^2)

Measures edge strength before non-maximum suppression and thresholding.

Pseudocode

1.Smooth the grayscale image with a Gaussian filter
2.Compute gradient magnitude and direction
3.Apply non-maximum suppression along gradient directions
4.Mark strong and weak edge pixels with two thresholds
5.Keep weak edges connected to strong edges

Canonical Example

Input: a grayscale image with a dark square on a light background
Output: thin edge pixels around the square boundary
Explanation: The square boundary creates strong gradients that survive suppression and hysteresis.

Implementation Notes

Choose Gaussian sigma and thresholds together.
Test non-maximum suppression on simple diagonal and vertical edges.
Keep border handling consistent between smoothing and gradient stages.

Common Pitfalls

Too little smoothing creates noisy edges; too much smoothing erases fine detail.
Poor high and low thresholds either break edges apart or keep too much noise.
Incorrect gradient-direction quantization makes non-maximum suppression jagged.