Package com.irurueta.geometry.refiners

Class DecomposedLinePlaneCorrespondencePinholeCameraRefiner

java.lang.Object

com.irurueta.geometry.refiners.Refiner<PinholeCamera>

com.irurueta.geometry.refiners.InliersDataRefiner<PinholeCamera>

com.irurueta.geometry.refiners.PairMatchesAndInliersDataRefiner<PinholeCamera,Plane,Line2D>

com.irurueta.geometry.refiners.PinholeCameraRefiner<Plane,Line2D>

com.irurueta.geometry.refiners.LinePlaneCorrespondencePinholeCameraRefiner

com.irurueta.geometry.refiners.DecomposedLinePlaneCorrespondencePinholeCameraRefiner

public class DecomposedLinePlaneCorrespondencePinholeCameraRefiner
extends LinePlaneCorrespondencePinholeCameraRefiner

A pinhole camera refiner using line/plane correspondences and the Powell algorithm to try to decrease overall error in LMSE terms among inlier samples by taking the decomposed parameters of a pinhole camera. Typically, this refiner is used by a robust estimator, however it can also be useful in some other situations.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
private class	DecomposedLinePlaneCorrespondencePinholeCameraRefiner.RefinementMultiDimensionFunctionEvaluatorListener	Listener for powell optimizer to minimize cost function during refinement.

Field Summary

Fields

Modifier and Type	Field	Description
private double	currentWeight	Current weight during refinement.
static final double	DEFAULT_MAX_SUGGESTION_WEIGHT	Default value for maximum suggestion weight.
static final double	DEFAULT_MIN_SUGGESTION_WEIGHT	Default value for minimum suggestion weight.
static final double	DEFAULT_SUGGESTION_WEIGHT_STEP	Default value for the step to increase suggestion weight.
private double	maxSuggestionWeight	Maximum suggestion weight.
private double	minSuggestionWeight	Minimum suggestion weight.
private static final int	REFINE_DIMS	Dimensions for refinement.
private PinholeCamera	refineCamera	Instance of a pinhole camera to be reused during refinement.
private double	suggestionWeightStep	Step to increase suggestion weight.

Fields inherited from class com.irurueta.geometry.refiners.PinholeCameraRefiner

DEFAULT_SUGGEST_ASPECT_RATIO_ENABLED, DEFAULT_SUGGEST_CENTER_ENABLED, DEFAULT_SUGGEST_HORIZONTAL_FOCAL_LENGTH_ENABLED, DEFAULT_SUGGEST_PRINCIPAL_POINT_ENABLED, DEFAULT_SUGGEST_ROTATION_ENABLED, DEFAULT_SUGGEST_SKEWNESS_VALUE_ENABLED, DEFAULT_SUGGEST_VERTICAL_FOCAL_LENGTH_ENABLED, DEFAULT_SUGGESTED_ASPECT_RATIO_VALUE, DEFAULT_SUGGESTED_SKEWNESS_VALUE, refinementStandardDeviation

Fields inherited from class com.irurueta.geometry.refiners.PairMatchesAndInliersDataRefiner

samples1, samples2

Fields inherited from class com.irurueta.geometry.refiners.InliersDataRefiner

inliers, numInliers, residuals

Fields inherited from class com.irurueta.geometry.refiners.Refiner

covariance, DEFAULT_KEEP_COVARIANCE, initialEstimation, keepCovariance, listener, locked

Constructor Summary

Constructors

Constructor	Description
DecomposedLinePlaneCorrespondencePinholeCameraRefiner()	Constructor.
DecomposedLinePlaneCorrespondencePinholeCameraRefiner(PinholeCamera initialEstimation, boolean keepCovariance, com.irurueta.numerical.robust.InliersData inliersData, List<Plane> samples1, List<Line2D> samples2, double refinementStandardDeviation)	Constructor.
DecomposedLinePlaneCorrespondencePinholeCameraRefiner(PinholeCamera initialEstimation, boolean keepCovariance, BitSet inliers, double[] residuals, int numInliers, List<Plane> samples1, List<Line2D> samples2, double refinementStandardDeviation)	Constructor.

Method Summary

Modifier and Type	Method	Description
private com.irurueta.algebra.Matrix	estimateCovarianceLevenbergMarquardt(PinholeCamera pinholeCamera, double weight)	Estimates covariance matrix for provided estimated and refined camera
double	getMaxSuggestionWeight()	Gets maximum suggestion weight.
double	getMinSuggestionWeight()	Gets minimum suggestion weight.
double	getSuggestionWeightStep()	Gets step to increase suggestion weight.
boolean	refine(PinholeCamera result)	Refines provided initial estimation.
private boolean	refinementStepPowell(com.irurueta.numerical.optimization.PowellMultiOptimizer optimizer, DecomposedLinePlaneCorrespondencePinholeCameraRefiner.RefinementMultiDimensionFunctionEvaluatorListener listener, double[] startPoint, double weight)	Computes one refinement step using Powell optimizer for a given weight on suggestion terms.
private boolean	refinePowell(PinholeCamera result)	Refines camera using Powell optimization to minimize a cost function consisting on the sum of squared projection residuals plus the suggestion residual for any suggested terms.
void	setMaxSuggestionWeight(double maxSuggestionWeight)	Sets maximum suggestion weight.
void	setMinMaxSuggestionWeight(double minSuggestionWeight, double maxSuggestionWeight)	Sets minimum and maximum suggestion weights.
void	setMinSuggestionWeight(double minSuggestionWeight)	Sets minimum suggestion weight.
void	setSuggestionWeightStep(double suggestionWeightStep)	Sets step to increase suggestion weight.

Methods inherited from class com.irurueta.geometry.refiners.LinePlaneCorrespondencePinholeCameraRefiner

residualLevenbergMarquardt, residualPowell

Methods inherited from class com.irurueta.geometry.refiners.PinholeCameraRefiner

cameraToParameters, getRefinementStandardDeviation, getSuggestedAspectRatioValue, getSuggestedCenterValue, getSuggestedHorizontalFocalLengthValue, getSuggestedPrincipalPointValue, getSuggestedRotationValue, getSuggestedSkewnessValue, getSuggestedVerticalFocalLengthValue, hasSuggestions, isSuggestAspectRatioEnabled, isSuggestCenterEnabled, isSuggestHorizontalFocalLengthEnabled, isSuggestPrincipalPointEnabled, isSuggestRotationEnabled, isSuggestSkewnessValueEnabled, isSuggestVerticalFocalLengthEnabled, parametersToCamera, refine, setRefinementStandardDeviation, setSuggestAspectRatioEnabled, setSuggestCenterEnabled, setSuggestedAspectRatioValue, setSuggestedCenterValue, setSuggestedHorizontalFocalLengthValue, setSuggestedPrincipalPointValue, setSuggestedRotationValue, setSuggestedSkewnessValue, setSuggestedVerticalFocalLengthValue, setSuggestHorizontalFocalLengthEnabled, setSuggestPrincipalPointEnabled, setSuggestRotationEnabled, setSuggestSkewnessValueEnabled, setSuggestVerticalFocalLengthEnabled, suggestionResidual

Methods inherited from class com.irurueta.geometry.refiners.PairMatchesAndInliersDataRefiner

getSamples1, getSamples2, isReady, setSamples1, setSamples2

Methods inherited from class com.irurueta.geometry.refiners.InliersDataRefiner

getInliers, getNumInliers, getResiduals, getTotalSamples, setInliers, setInliersData, setNumInliers, setResiduals

Methods inherited from class com.irurueta.geometry.refiners.Refiner

getCovariance, getInitialEstimation, getListener, isCovarianceKept, isLocked, setCovarianceKept, setInitialEstimation, setListener

Methods inherited from class java.lang.Object

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

Field Details

DEFAULT_MIN_SUGGESTION_WEIGHT

public static final double DEFAULT_MIN_SUGGESTION_WEIGHT

Default value for minimum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

See Also:

• Constant Field Values

DEFAULT_MAX_SUGGESTION_WEIGHT

public static final double DEFAULT_MAX_SUGGESTION_WEIGHT

Default value for maximum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

See Also:

• Constant Field Values

DEFAULT_SUGGESTION_WEIGHT_STEP

public static final double DEFAULT_SUGGESTION_WEIGHT_STEP

Default value for the step to increase suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

See Also:

• Constant Field Values

REFINE_DIMS

private static final int REFINE_DIMS

Dimensions for refinement.

See Also:

• Constant Field Values

minSuggestionWeight

private double minSuggestionWeight

Minimum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

maxSuggestionWeight

private double maxSuggestionWeight

Maximum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

suggestionWeightStep

private double suggestionWeightStep

Step to increase suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

refineCamera

private PinholeCamera refineCamera

Instance of a pinhole camera to be reused during refinement.

currentWeight

private double currentWeight

Current weight during refinement.

Constructor Details

DecomposedLinePlaneCorrespondencePinholeCameraRefiner

public DecomposedLinePlaneCorrespondencePinholeCameraRefiner()

Constructor.

DecomposedLinePlaneCorrespondencePinholeCameraRefiner

public DecomposedLinePlaneCorrespondencePinholeCameraRefiner(PinholeCamera initialEstimation,
 boolean keepCovariance,
 BitSet inliers,
 double[] residuals,
 int numInliers,
 List<Plane> samples1,
 List<Line2D> samples2,
 double refinementStandardDeviation)

Constructor.

Parameters:

initialEstimation - initial estimation to be set.

keepCovariance - true if covariance of estimation must be kept after refinement, false otherwise.

inliers - set indicating which of the provided matches are inliers.

residuals - residuals for matched samples.

numInliers - number of inliers on initial estimation.

samples1 - 1st set of paired samples.

samples2 - 2nd set of paired samples.

refinementStandardDeviation - standard deviation used for Levenberg-Marquardt fitting.

DecomposedLinePlaneCorrespondencePinholeCameraRefiner

public DecomposedLinePlaneCorrespondencePinholeCameraRefiner(PinholeCamera initialEstimation,
 boolean keepCovariance,
 com.irurueta.numerical.robust.InliersData inliersData,
 List<Plane> samples1,
 List<Line2D> samples2,
 double refinementStandardDeviation)

Constructor.

Parameters:

initialEstimation - initial estimation to be set.

keepCovariance - true if covariance of estimation must be kept after refinement, false otherwise.

inliersData - inlier data, typically obtained from a robust estimator.

samples1 - 1st set of paired samples.

samples2 - 2nd set of paired samples.

refinementStandardDeviation - standard deviation used for Levenberg-Marquardt fitting.

Method Details

getMinSuggestionWeight

public double getMinSuggestionWeight()

Gets minimum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Returns:

minimum suggestion weight.

setMinSuggestionWeight

public void setMinSuggestionWeight(double minSuggestionWeight)
                            throws LockedException

Sets minimum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Parameters:

minSuggestionWeight - minimum suggestion weight.

Throws:

LockedException - if estimator is locked.

getMaxSuggestionWeight

public double getMaxSuggestionWeight()

Gets maximum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Returns:

maximum suggestion weight.

setMaxSuggestionWeight

public void setMaxSuggestionWeight(double maxSuggestionWeight)
                            throws LockedException

Sets maximum suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Parameters:

maxSuggestionWeight - maximum suggestion weight.

Throws:

LockedException - if estimator is locked.

setMinMaxSuggestionWeight

public void setMinMaxSuggestionWeight(double minSuggestionWeight,
 double maxSuggestionWeight)
                               throws LockedException

Sets minimum and maximum suggestion weights. Suggestion weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Parameters:

minSuggestionWeight - minimum suggestion weight.

maxSuggestionWeight - maximum suggestion weight.

Throws:

LockedException - if estimator is locked.

IllegalArgumentException - if minimum suggestion weight is greater or equal than maximum value.

getSuggestionWeightStep

public double getSuggestionWeightStep()

Gets step to increase suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Returns:

step to increase suggestion weight.

setSuggestionWeightStep

public void setSuggestionWeightStep(double suggestionWeightStep)
                             throws LockedException

Sets step to increase suggestion weight. This weight is used to slowly draw original camera parameters into desired suggested values. Suggestion weight slowly increases each time Levenberg-Marquardt is used to find a solution so that the algorithm can converge into desired value. The faster the weights are increased the less likely that suggested values can be converged if they differ too much from the original ones.

Parameters:

suggestionWeightStep - step to increase suggestion weight.

Throws:

LockedException - if estimator is locked.

IllegalArgumentException - if provided step is negative or zero.

refine

public boolean refine(PinholeCamera result)
               throws NotReadyException,
LockedException

Refines provided initial estimation. This method always sets a value into provided result instance regardless of the fact that error has actually improved in LMSE terms or not.

Specified by:

refine in class Refiner<PinholeCamera>

Parameters:

result - instance where refined estimation will be stored.

Returns:

true if result improves (decreases) in LMSE terms respect to initial estimation, false if no improvement has been achieved.

Throws:

NotReadyException - if not enough input data has been provided.

LockedException - if estimator is locked because refinement is already in progress.

estimateCovarianceLevenbergMarquardt

private com.irurueta.algebra.Matrix estimateCovarianceLevenbergMarquardt(PinholeCamera pinholeCamera,
 double weight)

Estimates covariance matrix for provided estimated and refined camera

Parameters:

pinholeCamera - pinhole camera to estimate covariance for.

weight - weight for suggestion residual.

Returns:

estimated covariance or null if anything fails.

refinePowell

private boolean refinePowell(PinholeCamera result)

Refines camera using Powell optimization to minimize a cost function consisting on the sum of squared projection residuals plus the suggestion residual for any suggested terms.

Parameters:

result - instance where refined estimation will be stored.

Returns:

true if result improves (decreases) in LMSE terms respect to initial estimation, false if no improvement has been achieved.

refinementStepPowell

private boolean refinementStepPowell(com.irurueta.numerical.optimization.PowellMultiOptimizer optimizer,
 DecomposedLinePlaneCorrespondencePinholeCameraRefiner.RefinementMultiDimensionFunctionEvaluatorListener listener,
 double[] startPoint,
 double weight)
                              throws GeometryException,
com.irurueta.numerical.NumericalException

Computes one refinement step using Powell optimizer for a given weight on suggestion terms.

Parameters:

optimizer - Powell optimizer to be reused.

listener - Powell optimizer listener to be reused.

startPoint - starting point for powell optimization. This array is passed only for reuse purposes.

weight - suggestion terms weight.

Returns:

true if this refinement step decreased projection error in LMSE terms, false otherwise.

Throws:

GeometryException - if something failed.

com.irurueta.numerical.NumericalException - if something failed.