VID · Video tools

Edge Detection Video

Shareable link

Settings are written to the URL as you change them. Nothing differs from the defaults yet.

Reduce a clip to its outlines. Each frame is convolved with a 3×3 derivative kernel, and what remains is where brightness changes: object edges, texture, and text.

Operators

Sobel, Prewitt, and Scharr measure the rate of brightness change horizontally and vertically and combine the two. They differ only in how they weight the center row and column. Laplacian measures where that rate turns over, which gives thin lines on both sides of an edge.

Noise in motion

Sensor and compression noise change every frame, so an unsmoothed edge map crawls with speckle. Pre-Blur is the main fix. Sensitivity scales the gradient: raise it to bring faint edges up, lower it when texture drowns the outlines. A Binary Threshold gives clean line art that flickers only where edges are weak.

Uses

Edge-detected footage works as a source for rotoscoping references, motion graphics backgrounds, and computer vision demos. For a colored glow along the lines, use the neon outline video effect.

Frequently Asked Questions

Sobel for most footage: its built-in smoothing keeps noise down. Scharr is the most accurate on diagonal edges. Prewitt is noisier. Laplacian gives thinner, often doubled lines because it is a second derivative.

A derivative amplifies noise, and camera footage is noisy. Raise Pre-Blur, which smooths the frame before the kernel runs; its radius is a share of the short edge, so it behaves the same at every size.

Above 0 it turns the gradient into pure on or off lines: pixels above the threshold become lines and everything else the background.

It is copied into the export unchanged when the container can hold its codec, and re-encoded when it cannot. A note under the export button says which happened.

Explore Our Tools

Browse all tools