org.openimaj.video.tracking.klt

Class KLTTracker

java.lang.Object

org.openimaj.video.tracking.klt.KLTTracker

public class KLTTracker
extends Object

The KLT tracker

Author:

Stan Birchfield, Jonathon Hare (jsh2@ecs.soton.ac.uk)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	KLTTracker.SelectionMode Modes of operation for selecting features.

Field Summary

Fields

Modifier and Type	Field and Description
static int	KLT_LARGE_RESIDUE The residue was too large
static int	KLT_MAX_ITERATIONS The maximum number of iterations was exceeded
static int	KLT_NOT_FOUND The feature was not found
static int	KLT_OOB The feature was out of bouns
static int	KLT_SMALL_DET The determinant was too small
static int	KLT_TRACKED The feature was tracked
protected int	KLT_verbose

Constructor Summary

Constructors

Constructor and Description
KLTTracker(int nfeatures) Construct with the given target number of features.
KLTTracker(TrackingContext tc, FeatureList featurelist) Construct with the given context and feature list

Method Summary

Modifier and Type	Method and Description
FeatureList	getFeatureList()
TrackingContext	getTrackingContext()
boolean	isNorm()
void	replaceLostFeatures(FImage img) KLTReplaceLostFeatures Main routine, visible to the outside.
void	selectGoodFeatures(FImage img) KLTSelectGoodFeatures Main routine, visible to the outside.
void	setFeatureList(FeatureList featurelist) Set the tracking context
void	setNorm(boolean isNorm) Set whether input images are in [0,1] (true) or [0,255] (false).
void	setTrackingContext(TrackingContext tc) Set the tracking context
void	setVerbosity(int verbosity) KLTSetVerbosity
void	trackFeatures(FImage img1, FImage img2) KLTTrackFeatures Tracks feature points from one image to the next.
void	trackFeatures(FImage img1, FImage img2, PyramidSet pyr1, PyramidSet pyr2) KLTTrackFeatures Tracks feature points from one image to the next.

Methods inherited from class java.lang.Object

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

Field Detail

KLT_verbose

protected int KLT_verbose

KLT_TRACKED

public static final int KLT_TRACKED

The feature was tracked

See Also:

Constant Field Values

KLT_NOT_FOUND

public static final int KLT_NOT_FOUND

The feature was not found

See Also:

Constant Field Values

KLT_SMALL_DET

public static final int KLT_SMALL_DET

The determinant was too small

See Also:

Constant Field Values

KLT_MAX_ITERATIONS

public static final int KLT_MAX_ITERATIONS

The maximum number of iterations was exceeded

See Also:

Constant Field Values

KLT_OOB

public static final int KLT_OOB

The feature was out of bouns

See Also:

Constant Field Values

KLT_LARGE_RESIDUE

public static final int KLT_LARGE_RESIDUE

The residue was too large

See Also:

Constant Field Values

Constructor Detail

KLTTracker

public KLTTracker(int nfeatures)

Construct with the given target number of features.

Parameters:

nfeatures -

KLTTracker

public KLTTracker(TrackingContext tc,
                  FeatureList featurelist)

Construct with the given context and feature list

Parameters:

tc -

featurelist -

Method Detail

selectGoodFeatures

public void selectGoodFeatures(FImage img)

KLTSelectGoodFeatures Main routine, visible to the outside. Finds the good features in an image. INPUTS tc: Contains parameters used in computation (size of image, size of window, min distance b/w features, sigma to compute image gradients, # of features desired).

Parameters:

img - Pointer to the data of an image (probably unsigned chars). OUTPUTS features: List of features. The member nFeatures is computed.

replaceLostFeatures

public void replaceLostFeatures(FImage img)

KLTReplaceLostFeatures Main routine, visible to the outside. Replaces the lost features in an image. INPUTS tc: Contains parameters used in computation (size of image, size of window, min distance b/w features, sigma to compute image gradients, # of features desired).

Parameters:

img - Pointer to the data of an image (probably unsigned chars). OUTPUTS features: List of features. The member nFeatures is computed.

setVerbosity

public void setVerbosity(int verbosity)

KLTSetVerbosity

Parameters:

verbosity -

trackFeatures

public void trackFeatures(FImage img1,
                          FImage img2)

KLTTrackFeatures Tracks feature points from one image to the next.

Parameters:

img1 -

img2 -

trackFeatures

public void trackFeatures(FImage img1,
                          FImage img2,
                          PyramidSet pyr1,
                          PyramidSet pyr2)

KLTTrackFeatures Tracks feature points from one image to the next. Also takes in image pyramids and gradient pyramids for both images.

Parameters:

img1 -

img2 -

pyr1 -

pyr2 -

getTrackingContext

public TrackingContext getTrackingContext()

Returns:

the tracking context

setTrackingContext

public void setTrackingContext(TrackingContext tc)

Set the tracking context

Parameters:

tc -

getFeatureList

public FeatureList getFeatureList()

Returns:

the feature list

setFeatureList

public void setFeatureList(FeatureList featurelist)

Set the tracking context

Parameters:

featurelist -

isNorm

public boolean isNorm()

Returns:

true if input images are normalised in [0,1]; false if in [0, 255]

setNorm

public void setNorm(boolean isNorm)

Set whether input images are in [0,1] (true) or [0,255] (false).

Parameters:

isNorm -