public interface Model<M extends Model<M>> extends CoordinateTransform
CoordinateTransform whose parameters can be estimated through
a least-squares(like) fit.
Provides methods for generic optimization and model extraction algorithms.
Currently, the Random Sample Consensus \cite{FischlerB81}, a robust
regression method and the Iterative Closest Point Algorithm \cite{Zhang94}
are implemented.
BibTeX:
@article{FischlerB81,
author = {Martin A. Fischler and Robert C. Bolles},
title = {Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography},
journal = {Communications of the ACM},
volume = {24},
number = {6},
year = {1981},
pages = {381--395},
publisher = {ACM Press},
address = {New York, NY, USA},
issn = {0001-0782},
doi = {http://doi.acm.org/10.1145/358669.358692},
}
@article{Zhang94,
author = {{Zhengyou Zhang}},
title = {Iterative point matching for registration of free-form curves and surfaces},
journal = {International Journal of Computer Vision},
volume = {13},
number = {2},
month = {October},
year = {1994},
pages = {119--152},
}
| Modifier and Type | Method and Description |
|---|---|
boolean |
betterThan(M m)
"Less than" operater to make
Models comparable. |
M |
copy()
Clone the model.
|
<P extends PointMatch> |
filter(Collection<P> candidates,
Collection<P> inliers)
Call
filter(Collection, Collection, double) with maxTrust = 4 and minNumInliers = getMinNumMatches(). |
<P extends PointMatch> |
filter(Collection<P> candidates,
Collection<P> inliers,
double maxTrust)
Call
filter(Collection, Collection, double, int) with minNumInliers = getMinNumMatches(). |
<P extends PointMatch> |
filter(Collection<P> candidates,
Collection<P> inliers,
double maxTrust,
int minNumInliers)
Estimate the
Model and filter potential outliers by robust
iterative regression. |
<P extends PointMatch> |
filterRansac(List<P> candidates,
Collection<P> inliers,
int iterations,
double maxEpsilon,
double minInlierRatio)
Call
filterRansac(List, Collection, int, double, double, double)
with maxTrust = 4. |
<P extends PointMatch> |
filterRansac(List<P> candidates,
Collection<P> inliers,
int iterations,
double maxEpsilon,
double minInlierRatio,
double maxTrust)
Call
filterRansac(List, Collection, int, double, double, int, double)
with minNumInliers = getMinNumMatches(). |
<P extends PointMatch> |
filterRansac(List<P> candidates,
Collection<P> inliers,
int iterations,
double maxEpsilon,
double minInlierRatio,
int minNumInliers)
Call
filterRansac(List, Collection, int, double, double, int, double)
with maxTrust = 4. |
<P extends PointMatch> |
filterRansac(List<P> candidates,
Collection<P> inliers,
int iterations,
double maxEpsilon,
double minInlierRatio,
int minNumInliers,
double maxTrust)
|
<P extends PointMatch> |
fit(Collection<P> matches)
Fit the
Model to a set of data points minimizing the global
transfer error. |
void |
fit(double[][] p,
double[][] q,
double[] w)
Fit the
Model to a set of data points minimizing the global
transfer error. |
void |
fit(float[][] p,
float[][] q,
float[] w)
Fit the
Model to a set of data points minimizing the global
transfer error. |
double |
getCost() |
int |
getMinNumMatches() |
Collection<PointMatch> |
icp(List<Point> p,
List<Point> q)
Estimate the best model in terms of the Iterative Closest Point
Algorithm \cite{Zhang94} for matching two point clouds into each other.
|
<P extends PointMatch> |
localSmoothnessFilter(Collection<P> candidates,
Collection<P> inliers,
double sigma,
double maxEpsilon,
double maxTrust)
|
<P extends PointMatch> |
ransac(List<P> candidates,
Collection<P> inliers,
int iterations,
double epsilon,
double minInlierRatio)
Call
ransac(List, Collection, int, double, double, int) with
minNumInliers = getMinNumMatches(). |
<P extends PointMatch> |
ransac(List<P> candidates,
Collection<P> inliers,
int iterations,
double epsilon,
double minInlierRatio,
int minNumInliers)
Find the
Model of a set of PointMatch candidates
containing a high number of outliers using
RANSAC
\citet[{FischlerB81}. |
void |
set(M m)
Set the model to m
|
void |
setCost(double c) |
<P extends PointMatch> |
test(Collection<P> candidates,
Collection<P> inliers,
double epsilon,
double minInlierRatio)
Call
test(Collection, Collection, double, double, int) with
minNumInliers = getMinNumMatches(). |
<P extends PointMatch> |
test(Collection<P> candidates,
Collection<P> inliers,
double epsilon,
double minInlierRatio,
int minNumInliers)
Test the
Model for a set of PointMatch candidates. |
apply, applyInPlaceint getMinNumMatches()
PointMatches required
to solve the model.double getCost()
void setCost(double c)
boolean betterThan(M m)
Models comparable.m - cost < 0.0, otherwise true if
this.getCost() is smaller than
m.getCost()void fit(float[][] p,
float[][] q,
float[] w)
throws NotEnoughDataPointsException,
IllDefinedDataPointsException
Fit the Model to a set of data points minimizing the global
transfer error. This is assumed to be implemented as a weighted least
squares minimization.
This is a lower level version of fit(Collection) for
optimal memory efficiency.
The estimated model transfers p to q.
n-dimensional points are passed as an n-dimensional array of floats, e.g. four 2d points as:
float[][]{
{x1, x2, x3, x4},
{y1, y2, y3, y4} }
p - source pointsq - target pointsw - weightsNotEnoughDataPointsException - if not enough data points
were availableIllDefinedDataPointsException - if the set of data points is
inappropriate to solve the Modelvoid fit(double[][] p,
double[][] q,
double[] w)
throws NotEnoughDataPointsException,
IllDefinedDataPointsException
Fit the Model to a set of data points minimizing the global
transfer error. This is assumed to be implemented as a weighted least
squares minimization.
This is a lower level version of fit(Collection) for
optimal memory efficiency.
The estimated model transfers p to q.
n-dimensional points are passed as an n-dimensional array of doubles, e.g. four 2d points as:
double[][]{
{x1, x2, x3, x4},
{y1, y2, y3, y4} }
p - source pointsq - target pointsw - weightsNotEnoughDataPointsException - if not enough data points
were availableIllDefinedDataPointsException - if the set of data points is
inappropriate to solve the Model<P extends PointMatch> void fit(Collection<P> matches) throws NotEnoughDataPointsException, IllDefinedDataPointsException
Model to a set of data points minimizing the global
transfer error. This is assumed to be implemented as a weighted least
squares minimization. Use ransac
and/ or filter(java.util.Collection<P>, java.util.Collection<P>, double, int) to remove outliers from your data points.
The estimated model transfers match.p1.local to match.p2.world.
matches - set of point correpondencesNotEnoughDataPointsException - if matches does not contain
enough data pointsIllDefinedDataPointsException - if the set of data points is
inappropriate to solve the Model<P extends PointMatch> boolean test(Collection<P> candidates, Collection<P> inliers, double epsilon, double minInlierRatio, int minNumInliers)
Model for a set of PointMatch candidates.
Return true if the number of inliers / number of candidates is larger
than or equal to min_inlier_ratio, otherwise false.
Clears inliers and fills it with the fitting subset of candidates.
Sets cost = 1.0 - |inliers| / |candidates|.
candidates - set of point correspondence candidatesinliers - set of point correspondences that fit the modelepsilon - maximal allowed transfer errorminInlierRatio - minimal ratio |inliers| / |candidates| (0.0 => 0%, 1.0 => 100%)minNumInliers - minimally required absolute number of inliers<P extends PointMatch> boolean test(Collection<P> candidates, Collection<P> inliers, double epsilon, double minInlierRatio)
test(Collection, Collection, double, double, int) with
minNumInliers = getMinNumMatches().<P extends PointMatch> boolean filter(Collection<P> candidates, Collection<P> inliers, double maxTrust, int minNumInliers) throws NotEnoughDataPointsException
Model and filter potential outliers by robust
iterative regression.
This method performs well on data sets with low amount of outliers. If
you have many outliers, you can filter those with a `tolerant' RANSAC
first as done in filterRansac(java.util.List<P>, java.util.Collection<P>, int, double, double, int, double).
Sets cost to the average point transfer error.
candidates - Candidate data points eventually inluding some outliersinliers - Remaining after the robust regression filtermaxTrust - reject candidates with a cost larger than
maxTrust * median costminNumInliers - minimally required absolute number of inliersModel could be estimated and inliers is not
empty, false otherwise. If false, Model remains unchanged.NotEnoughDataPointsException<P extends PointMatch> boolean filter(Collection<P> candidates, Collection<P> inliers, double maxTrust) throws NotEnoughDataPointsException
filter(Collection, Collection, double, int) with minNumInliers = getMinNumMatches().NotEnoughDataPointsException<P extends PointMatch> boolean filter(Collection<P> candidates, Collection<P> inliers) throws NotEnoughDataPointsException
filter(Collection, Collection, double) with maxTrust = 4 and minNumInliers = getMinNumMatches().NotEnoughDataPointsException<P extends PointMatch> boolean ransac(List<P> candidates, Collection<P> inliers, int iterations, double epsilon, double minInlierRatio, int minNumInliers) throws NotEnoughDataPointsException
Model of a set of PointMatch candidates
containing a high number of outliers using
RANSAC
\citet[{FischlerB81}.candidates - candidate data points inluding (many) outliersinliers - remaining candidates after RANSACiterations - number of iterationsepsilon - maximal allowed transfer errorminInlierRatio - minimal number of inliers to number of
candidatesminNumInliers - minimally required absolute number of inliersModel could be estimated and inliers is not
empty, false otherwise. If false, Model remains unchanged.NotEnoughDataPointsException<P extends PointMatch> boolean ransac(List<P> candidates, Collection<P> inliers, int iterations, double epsilon, double minInlierRatio) throws NotEnoughDataPointsException
ransac(List, Collection, int, double, double, int) with
minNumInliers = getMinNumMatches().NotEnoughDataPointsException<P extends PointMatch> boolean filterRansac(List<P> candidates, Collection<P> inliers, int iterations, double maxEpsilon, double minInlierRatio, int minNumInliers, double maxTrust) throws NotEnoughDataPointsException
Model from a set with many outliers by first
filtering the worst outliers with RANSAC
\citet[{FischlerB81} and filter potential outliers by robust iterative
regression.candidates - candidate data points inluding (many) outliersinliers - remaining candidates after RANSACiterations - number of iterationsmaxEpsilon - maximal allowed transfer errorminInlierRatio - minimal number of inliers to number of
candidatesminNumInliers - minimally required absolute number of inliersmaxTrust - reject candidates with a cost larger than
maxTrust * median costModel could be estimated and inliers is not
empty, false otherwise. If false, Model remains unchanged.NotEnoughDataPointsException<P extends PointMatch> boolean filterRansac(List<P> candidates, Collection<P> inliers, int iterations, double maxEpsilon, double minInlierRatio, int minNumInliers) throws NotEnoughDataPointsException
filterRansac(List, Collection, int, double, double, int, double)
with maxTrust = 4.NotEnoughDataPointsException<P extends PointMatch> boolean filterRansac(List<P> candidates, Collection<P> inliers, int iterations, double maxEpsilon, double minInlierRatio, double maxTrust) throws NotEnoughDataPointsException
filterRansac(List, Collection, int, double, double, int, double)
with minNumInliers = getMinNumMatches().NotEnoughDataPointsException<P extends PointMatch> boolean filterRansac(List<P> candidates, Collection<P> inliers, int iterations, double maxEpsilon, double minInlierRatio) throws NotEnoughDataPointsException
filterRansac(List, Collection, int, double, double, double)
with maxTrust = 4.NotEnoughDataPointsException<P extends PointMatch> boolean localSmoothnessFilter(Collection<P> candidates, Collection<P> inliers, double sigma, double maxEpsilon, double maxTrust) throws NotEnoughDataPointsException, IllDefinedDataPointsException
Collection of PointMatches by the
smoothness of their support for the given Model. Smoothness
means that each PointMatch agrees with a locally weighted least
squares fit of this Model up to a given maximal transfer error
or up to a given multiply of the local mean transfer error. Locally
weighted means that all PointMatches contribute to
the fit weighted by their Euclidean distance to the candidate.candidates - inliers - sigma - maxEpsilon - maxTrust - NotEnoughDataPointsExceptionIllDefinedDataPointsExceptionCollection<PointMatch> icp(List<Point> p, List<Point> q)
p -> q
p - sourceq - targetvoid set(M m)
m - M copy()
Copyright © 2015–2021 Fiji. All rights reserved.