estimateStereoRectification

Uncalibrated stereo rectification

Since R2022b

Syntax

[tform1,tform2] = estimateStereoRectification(F,inlierPoints1,inlierPoints2,imageSize)

Description

[tform1,tform2] = estimateStereoRectification(F,inlierPoints1,inlierPoints2,imageSize) returns projective transformations for rectifying stereo images. This function does not require intrinsic or extrinsic camera parameters.

Examples

Estimate and Apply Stereo Rectification

Load the stereo images and feature points which are already matched.

I1 = imread("yellowstone_left.png");
I2 = imread("yellowstone_right.png");
load yellowstone_inlier_points;

Display point correspondences. Notice that the matching points are in different rows, indicating that the stereo pair is not rectified.

showMatchedFeatures(I1,I2,inlier_points1,inlier_points2,"montage");
title("Original Images and Matching Feature Points");

Calculate the fundamental matrix from the corresponding points.

f = estimateFundamentalMatrix(inlier_points1,inlier_points2, ...
    "Method","Norm8Point");

Calculate the rectification transformations.

[tform1,tform2] = estimateStereoRectification(f,inlier_points1,...
    inlier_points2,size(I2));

Rectify the stereo images using projective transformations tform1 and tform2.

[I1Rect,I2Rect] = rectifyStereoImages(I1,I2,tform1,tform2);

Display the stereo anaglyph, which can also be viewed with 3-D glasses.

figure
imshow(stereoAnaglyph(I1Rect,I2Rect))

Input Arguments

F — Fundamental matrix

3-by-3 matrix

Fundamental matrix for the stereo images, specified as a 3-by-3 matrix. The fundamental matrix satisfies this criteria for P₁, a point in image 1, and P₂, a corresponding point in image 2:

[P₂,1] * F * [P₁,1]' = 0

Data Types: single | double

inlierPoints1 — Coordinates of points in image 1

M-by-2 matrix | point feature object

Coordinates of points in image 1, specified as an M-by-2 matrix of M number of [x y] coordinates, or as one of the point feature objects described in Point Feature Types.

inlierPoints2 — Coordinates of corresponding points in image 2

M-by-2 matrix | point feature object

Coordinates of corresponding points in image 2, specified as an M-by-2 matrix of M number of [x y] coordinates, or as one of the point feature objects described in Point Feature Types.

imageSize — Size of image 2

numeric vector

Size of image 2, specified as a numeric vector in the format returned by the size function.

Output Arguments

tform1 — Projective transformation 1

projtform2d object

Projective transformation 1 describing the projective transformations for input image 1, returned as a projtform2d object.

tform2 — Projective transformation 2

projtform2d object

Projective transformation 2 describing the projective transformations for input image 2, returned as a projtform2d object.

Tips

• Applying the output uncalibrated rectification of tform1 (or tform2) to image 1 (or image 2) can result in an undesired distortion if an epipole exists within the image. You can check for an epipole within an image by applying the isEpipoleInImage function.

References

[1] Hartley, Richard, and Andrew Zisserman. Multiple View Geometry in Computer Vision. 2nd ed. Cambridge, UK ; New York: Cambridge University Press, 2003.

[2] Pollefeys, M., Koch, R., and Van Gool, L.. A Simple and Efficient Rectification Method for General Motion. Proceedings of the Seventh IEEE International Conference on Computer Vision. Volume 1, pages 496-501. 1999. DOI:10.1109/ICCV.1999.791262.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.

Version History

Introduced in R2022b

R2022b: Recommended over estimateUncalibratedRectification

Starting in R2022b, many Computer Vision Toolbox™ functions create and perform geometric transformations using the premultiply convention. However, the estimateUncalibratedRectification function uses the postmultiply convention. Although there are no plans to remove estimateUncalibratedRectification at this time, you can streamline your geometric transformation workflows by switching to the estimateStereoRectification function, which supports the premultiply convention. For more information, see Migrate Geometric Transformations to Premultiply Convention.