Skip to content
opencv-wasm

Affine transformations

Algorithm atlasGeometry

Map points with a linear transform plus translation, preserving straight lines and parallelism.

Try it on your image ↓

VISUAL WALKTHROUGHGeometry
REFERENCE INPUTSource coordinate frame

This example applies a known rotation, scale and translation to the source.

RESULTResampled output

The output preserves straight lines under this affine transform. Locations outside the source need an explicit border rule.

STEP 03 / 03

Resample the source and fill uncovered borders.

Illustrative example. The stages explain the method; they are not a live OpenCV execution.

Try it on an image

YOUR IMAGE · REAL OPENCV

Experiment at pixel level

Open full lab ↗

Rotate and scale around the image centre with a constant black border.

The engine loads on your first run. Your images stay in this browser.

Input448 × 320
OutputWaiting for a result

Scroll over either image to zoom at the pointer. Use the scrollbars to pan both views over the same relative area. Zoom is relative to the input; pixel coordinates belong to each image. Warps can change scene correspondence.

Pixel inspector RGBA · native values · matched scale · 9 × 9 output pixels
Hover to inspect. Click to pin a pixel.
Input 
Select a pixel

Output 
Select a pixel

Sample models and licenses

When to use it

Align images under rotation, scale, shear and translation, or rotate around a chosen centre.

How it works

  1. 01Estimate or construct a 2×3 transformation matrix.
  2. 02Map the output grid back to the input grid.
  3. 03Resample the source and fill uncovered borders.

[x′,y′]ᵀ = A[x,y]ᵀ + t

What to tune

WARP_INVERSE_MAP declares the direction of a supplied warp; dsize chooses the output canvas.

Where it breaks down

Affine models cannot describe general perspective distortion. Robust estimation needs enough well-distributed correspondences.

TypeScript API

Open an entry for its exact overloads, parameter descriptions, result ownership and pinned upstream source.

warpAffineReference →getAffineTransformReference →getRotationMatrix2DReference →estimateAffine2DReference →estimateAffinePartial2DReference →

All of these calls execute on the CPU. Native objects need explicit disposal. See matrices and ownership and build compatibility.

Image resizing

Sample the source on a new pixel grid at a different resolution.

Arbitrary remapping

Provide a source coordinate for every destination pixel.

Homography and RANSAC

Estimate a projective mapping between two views of a plane, rejecting inconsistent point matches.

Polar and log-polar warps

Re-express an image using angle and radius around a chosen centre.