C# Class Accord.Imaging.RansacHomographyEstimator

RANSAC Robust Homography Matrix Estimator.

Fitting a homography using RANSAC is pretty straightforward. Being a iterative method, in a single iteration a random sample of four correspondences is selected from the given correspondence points and a homography H is then computed from those points.

The original points are then transformed using this homography and their distances to where those transforms should be is then computed and matching points can classified as inliers and non-matching points as outliers.

After a given number of iterations, the iteration which produced the largest number of inliers is then selected as the best estimation for H.

References: E. Dubrofsky. Homography Estimation. Master thesis. Available on: http://www.cs.ubc.ca/~dubroe/courses/MastersEssay.pdf

Show file Open project: accord-net/framework Class Usage Examples

Public Methods

Method	Description
Estimate ( Accord points ) : MatrixH	Matches two sets of points using RANSAC.
Estimate ( Accord points1, Accord points2 ) : MatrixH	Matches two sets of points using RANSAC.
Estimate ( IntPoint points ) : MatrixH	Matches two sets of points using RANSAC.
Estimate ( IntPoint points1, IntPoint points2 ) : MatrixH	Matches two sets of points using RANSAC.
Estimate ( PointF points ) : MatrixH	Matches two sets of points using RANSAC.
Estimate ( PointF points1, PointF points2 ) : MatrixH	Matches two sets of points using RANSAC.
RansacHomographyEstimator ( double threshold, double probability ) : System	Creates a new RANSAC homography estimator.