Module imagingbook.calibrate.lib

Package imagingbook.calibration.zhang

Class HomographyEstimator

java.lang.Object

imagingbook.calibration.zhang.HomographyEstimator

public class HomographyEstimator
extends Object

This class defines methods for estimating the homography (projective) transformation between pairs of 2D point sets.

Field Summary

Fields

Modifier and Type	Field	Description
static int	MaxLmEvaluations
static int	MaxLmIterations

Constructor Summary

Constructors

Constructor	Description
HomographyEstimator()
HomographyEstimator(boolean normalizePointCoordinates, boolean doNonlinearRefinement)

Method Summary

Modifier and Type	Method	Description
RealMatrix[]	estimateHomographies(imagingbook.common.geometry.basic.Pnt2d[] modelPts, imagingbook.common.geometry.basic.Pnt2d[][] obsPoints)	Estimates the homographies between a fixed set of 2D model points and multiple observations (image point sets).
RealMatrix	estimateHomography(imagingbook.common.geometry.basic.Pnt2d[] ptsA, imagingbook.common.geometry.basic.Pnt2d[] ptsB)	Estimates the homography (projective) transformation from two given 2D point sets.
protected MultivariateMatrixFunction	getJacobianFunction(imagingbook.common.geometry.basic.Pnt2d[] X)
RealMatrix	refineHomography(RealMatrix Hinit, imagingbook.common.geometry.basic.Pnt2d[] pntsA, imagingbook.common.geometry.basic.Pnt2d[] pntsB)	Refines the initial homography by non-linear (Levenberg-Marquart) optimization.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

MaxLmEvaluations

public static int MaxLmEvaluations

MaxLmIterations

public static int MaxLmIterations

Constructor Details

HomographyEstimator

public HomographyEstimator()

HomographyEstimator

public HomographyEstimator(boolean normalizePointCoordinates,
 boolean doNonlinearRefinement)

Method Details

estimateHomographies

public RealMatrix[] estimateHomographies(imagingbook.common.geometry.basic.Pnt2d[] modelPts,
 imagingbook.common.geometry.basic.Pnt2d[][] obsPoints)

Estimates the homographies between a fixed set of 2D model points and multiple observations (image point sets). The correspondence between the points is assumed to be known.

Parameters:

modelPts - a sequence of 2D points on the model (calibration target)

obsPoints - a sequence 2D image point sets (one set per view).

Returns:

the sequence of estimated homographies (3 x 3 matrices), one for each view

estimateHomography

public RealMatrix estimateHomography(imagingbook.common.geometry.basic.Pnt2d[] ptsA,
 imagingbook.common.geometry.basic.Pnt2d[] ptsB)

Estimates the homography (projective) transformation from two given 2D point sets. The correspondence between the points is assumed to be known.

Parameters:

ptsA - the 1st sequence of 2D points

ptsB - the 2nd sequence of 2D points

Returns:

the estimated homography (3 x 3 matrix)

refineHomography

public RealMatrix refineHomography(RealMatrix Hinit,
 imagingbook.common.geometry.basic.Pnt2d[] pntsA,
 imagingbook.common.geometry.basic.Pnt2d[] pntsB)

Refines the initial homography by non-linear (Levenberg-Marquart) optimization.

Parameters:

Hinit - the initial (estimated) homography matrix

pntsA - the 1st sequence of 2D points

pntsB - the 2nd sequence of 2D points

Returns:

the refined homography matrix

getJacobianFunction

protected MultivariateMatrixFunction getJacobianFunction(imagingbook.common.geometry.basic.Pnt2d[] X)