Main Content

estimateEssentialMatrix

Estimate essential matrix from corresponding points in a pair of images

collapse all in page

Syntax

E = estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics)
E = estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics1,intrinsics2)
[E,inliersIndex] = estimateEssentialMatrix(___)
[E,inliersIndex,status] = estimateEssentialMatrix(___)
[___] = estimateEssentialMatrix(___,Name=Value)

Description

E = estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics) returns the 3-by-3 essential matrix, E, using the M-estimator sample consensus (MSAC) algorithm. The input points can be M-by-2 matrices of M number of [x,y] coordinates, or a KAZEPoints , SIFTPoints,SURFPoints, MSERRegions, BRISKPoints, or cornerPoints object. The instinsics object contains the intrinsic parameters of the camera used to take the images.

example

E = estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics1,intrinsics2) returns the essential matrix relating two images taken by different cameras. intrinsics1 and intrinsics2 are cameraIntrinsics objects containing the intrinsic parameters of camera 1 and camera 2 respectively.

[E,inliersIndex] = estimateEssentialMatrix(___) additionally returns an M-by-1 logical vector, inliersIndex, used to compute the essential matrix. The function sets the elements of the vector to true when the corresponding point was used to compute the fundamental matrix. The elements are set to false if they are not used.

[E,inliersIndex,status] = estimateEssentialMatrix(___) additionally returns a status code to indicate the validity of points.

[___] = estimateEssentialMatrix(___,Name=Value) specifies options using one or more name-value arguments in addition to any combination of arguments from previous syntaxes. For example, (MaxNumTrials=500) sets the maximum number of random trials for finding outliers to 500.

Examples

collapse all

Open Live Script

Load a MAT file containing precomputed camera parameters into the workspace.

load upToScaleReconstructionCameraParameters.mat

Read and undistort two images.

imageDir = fullfile(toolboxdir("vision"),"visiondata",...
    "upToScaleReconstructionImages");
images = imageDatastore(imageDir);
I1 = undistortImage(readimage(images,1),cameraParams);
I2 = undistortImage(readimage(images,2),cameraParams);
I1gray = im2gray(I1);
I2gray = im2gray(I2);

Detect feature points each image.

imagePoints1 = detectSURFFeatures(I1gray);
imagePoints2 = detectSURFFeatures(I2gray);

Extract feature descriptors from each image.

features1 = extractFeatures(I1gray,imagePoints1,Upright=true);
features2 = extractFeatures(I2gray,imagePoints2,Upright=true);

Match features across the images.

indexPairs = matchFeatures(features1,features2);
matchedPoints1 = imagePoints1(indexPairs(:,1));
matchedPoints2 = imagePoints2(indexPairs(:,2));
figure
showMatchedFeatures(I1,I2,matchedPoints1,matchedPoints2);
title("Putative Matches")

Figure contains an axes object. The hidden axes object with title Putative Matches contains 4 objects of type image, line. One or more of the lines displays its values using only markers

Estimate the essential matrix.

[E,inliers] = estimateEssentialMatrix(matchedPoints1,matchedPoints2,cameraParams);

Display the inlier matches.

inlierPoints1 = matchedPoints1(inliers);
inlierPoints2 = matchedPoints2(inliers);
figure
showMatchedFeatures(I1,I2,inlierPoints1,inlierPoints2);
title("Inlier Matches")

Figure contains an axes object. The hidden axes object with title Inlier Matches contains 4 objects of type image, line. One or more of the lines displays its values using only markers

Input Arguments

collapse all

Coordinates of corresponding points in image 1, specified as an M-by-2 matrix of M of [x,y] coordinates, or as a KAZEPoints, SIFTPoints, SURFPoints, BRISKPoints, MSERRegions, or cornerPoints object. The matchedPoints1 input must contain at least five points, which are putatively matched by using a function such as matchFeatures.

Coordinates of corresponding points in image 1, specified as an M-by-2 matrix of M of [x,y] coordinates, or as a KAZEPoints, SIFTPoints, SURFPoints, MSERRegions, BRISKPoints,or cornerPoints object. The matchedPoints2 input must contain at least five points, which are putatively matched by using a function such as matchFeatures.

Camera intrinsic parameters, specified as a cameraIntrinsics object.

Camera intrinsic parameters, specified as a cameraIntrinsics object.

Camera intrinsic parameters, specified as a cameraIntrinsics object.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Example: estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics,MaxNumTrials=500) sets the maximum number of random trials for finding outliers to 500.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: estimateEssentialMatrix(matchedPoints1,matchedPoints2,intrinsics,"MaxNumTrials",500) sets the maximum number of random trials for finding outliers to 500.

Maximum number of random trials for finding outliers, specified as the comma-separated pair consisting of 'MaxNumTrials' and a positive integer. The actual number of trials depends on matchedPoints1, matchedPoints2, and the value of the Confidence parameter. To optimize speed and accuracy, select the number of random trials.

Desired confidence for finding the maximum number of inliers, specified as the comma-separated pair consisting of 'Confidence' and a percentage scalar value in the range (0,100). Increasing this value improves the robustness of the output but increases the amount of computations.

Sampson distance threshold, specified as the comma-separated pair consisting of 'MaxDistance' and a scalar value. The function uses the threshold to find outliers returned in pixels squared. The Sampson distance is a first-order approximation of the squared geometric distance between a point and the epipolar line. Increase this value to make the algorithm converge faster, but this can also adversely affect the accuracy of the result.

Output Arguments

collapse all

Essential matrix, returned as a 3-by-3 matrix that is computed from the points in the matchedPoints1 and matchedPoints2 inputs with known camera intrinsics.

[P21]*EssentialMatrix*[P11]'=0

The P1 point in image 1, in normalized image coordinates, corresponds to the , P2 point in image 2.

In computer vision, the essential matrix is a 3-by-3 matrix which relates corresponding points in stereo images which are in normalized image coordinates. When two cameras view a 3-D scene from two distinct positions, the geometric relations between the 3-D points and their projections onto the 2-D images lead to constraints between image points. The two images of the same scene are related by epipolar geometry.

Data Types: double

Inliers index, returned as an M-by-1 logical index vector. An element set to true indicates that the corresponding indexed matched points in matchedPoints1 and matchedPoints2 were used to compute the essential matrix. An element set to false means the indexed points were not used for the computation.

Data Types: logical

Status code, returned as one of the following possible values:

statusValue
0:No error.
1:matchedPoints1 and matchedPoints2 do not contain enough points. At least five points are required.
2:Not enough inliers found. A least five inliers are required.

Data Types: int32

Tips

Use estimateEssentialMatrix when you know the camera intrinsics. You can obtain the intrinsics using the Camera Calibrator app. Otherwise, you can use the estimateFundamentalMatrix function that does not require camera intrinsics. The fundamental matrix cannot be estimated from coplanar world points.

References

[1] Kukelova, Z., M. Bujnak, and T. Pajdla Polynomial Eigenvalue Solutions to the 5-pt and 6-pt Relative Pose Problems. Leeds, UK: BMVC, 2008.

[2] Nister, D.. “An Efficient Solution to the Five-Point Relative Pose Problem.” IEEE Transactions on Pattern Analysis and Machine Intelligence.Volume 26, Issue 6, June 2004.

[3] Torr, P. H. S., and A. Zisserman. “MLESAC: A New Robust Estimator with Application to Estimating Image Geometry.” Computer Vision and Image Understanding. Volume 78, Issue 1, April 2000, pp. 138-156.

Extended Capabilities

Version History

Introduced in R2016b

expand all

See Also

Apps

Functions

Topics