boofcv.alg.geo

Class LowLevelMultiViewOps

java.lang.Object

boofcv.alg.geo.LowLevelMultiViewOps

public class LowLevelMultiViewOps
extends java.lang.Object

Lists of operations used by various multi-view algorithms, but not of use to the typical user.

Constructor Summary

Constructors

Constructor and Description
LowLevelMultiViewOps()

Method Summary

Modifier and Type	Method and Description
static void	applyPixelNormalization(org.ejml.data.DenseMatrix64F N, Point2D_F64 orig, Point2D_F64 normed) Given the normalization matrix computed from computeNormalization(java.util.List, org.ejml.data.DenseMatrix64F) normalize the point.
static void	computeNormalization(java.util.List<AssociatedPair> points, org.ejml.data.DenseMatrix64F N1, org.ejml.data.DenseMatrix64F N2) Computes two normalization matrices for each set of point correspondences in the list of AssociatedPair.
static void	computeNormalization(java.util.List<AssociatedTriple> points, org.ejml.data.DenseMatrix64F N1, org.ejml.data.DenseMatrix64F N2, org.ejml.data.DenseMatrix64F N3) Computes three normalization matrices for each set of point correspondences in the list of AssociatedTriple.
static void	computeNormalization(java.util.List<Point2D_F64> points, org.ejml.data.DenseMatrix64F N) Computes a normalization matrix used to reduce numerical errors inside of linear estimation algorithms.

Methods inherited from class java.lang.Object

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

Constructor Detail

LowLevelMultiViewOps

public LowLevelMultiViewOps()

Method Detail

computeNormalization

public static void computeNormalization(java.util.List<Point2D_F64> points,
                                        org.ejml.data.DenseMatrix64F N)

Computes a normalization matrix used to reduce numerical errors inside of linear estimation algorithms. Pixels coordinates are poorly scaled for linear algebra operations resulting in excessive numerical error. This function computes a transform which will minimize numerical error by properly scaling the pixels.

 N = [ 1/σ_x     0      -μ_x/σ_x ]
     [    0   1/σ_y 0   -μ_y/σ_y ]
     [    0      0          1    ]
 

Y. Ma, S. Soatto, J. Kosecka, and S. S. Sastry, "An Invitation to 3-D Vision" Springer-Verlad, 2004

Parameters:

points - Input: List of observed points. Not modified.

N - Output: 3x3 normalization matrix for first set of points. Modified.

computeNormalization

public static void computeNormalization(java.util.List<AssociatedPair> points,
                                        org.ejml.data.DenseMatrix64F N1,
                                        org.ejml.data.DenseMatrix64F N2)

Computes two normalization matrices for each set of point correspondences in the list of AssociatedPair. Same as computeNormalization(java.util.List, org.ejml.data.DenseMatrix64F), but for two views.

Parameters:

points - Input: List of observed points that are to be normalized. Not modified.

N1 - Output: 3x3 normalization matrix for first set of points. Modified.

N2 - Output: 3x3 normalization matrix for second set of points. Modified.

computeNormalization

public static void computeNormalization(java.util.List<AssociatedTriple> points,
                                        org.ejml.data.DenseMatrix64F N1,
                                        org.ejml.data.DenseMatrix64F N2,
                                        org.ejml.data.DenseMatrix64F N3)

Computes three normalization matrices for each set of point correspondences in the list of AssociatedTriple. Same as computeNormalization(java.util.List, org.ejml.data.DenseMatrix64F), but for three views.

Parameters:

points - Input: List of observed points that are to be normalized. Not modified.

N1 - Output: 3x3 normalization matrix for first set of points. Modified.

N2 - Output: 3x3 normalization matrix for second set of points. Modified.

N3 - Output: 3x3 normalization matrix for third set of points. Modified.

applyPixelNormalization

public static void applyPixelNormalization(org.ejml.data.DenseMatrix64F N,
                                           Point2D_F64 orig,
                                           Point2D_F64 normed)

Given the normalization matrix computed from computeNormalization(java.util.List, org.ejml.data.DenseMatrix64F) normalize the point.

Parameters:

N - Normalization matrix.

orig - Unnormalized coordinate in pixels.

normed - Normalized coordinate.