segmentImageMerge2 Class Reference

Collaboration diagram for segmentImageMerge2:

Public Member Functions
	segmentImageMerge2 (int instances)
	default constructor, also sets up number of tracker instances
void	resetAll (int instances)
void	SIMsetCircleColor (int r, int g, int b, int instance)
	set the tracking color for mass circle
void	SIMsetBoxColor (int r, int g, int b, int instance)
void	SIMsetTrackColor (float h, float hstd, float s, float sstd, float v, float vstd, int instance, bool adapt, int avg)
	set default tracking color values
void	SIMsetAdapt (float ha, bool haSTD, float sa, bool saSTD, float va, bool vaSTD, int instance, bool useCluster)
	set initial adaptive color thresholds for tracking
void	SIMsetAdaptBound (float Hupper, float Hlower, float Supper, float Slower, float Vupper, float Vlower, int instance)
	set up hard bounds for color adaptation
void	SIMsetCameraPosition (float pan, float tilt, int instance, bool stats=false)
	tell this method the pan and tilt postition the camera is moving to
void	SIMsetFrame (int x1, int y1, int x2, int y2, int realX, int realY, int instance)
	set frame size for color tracker
void	SIMtrackImage (Image< PixRGB< byte > > input, Image< PixRGB< byte > > image, int instance, Image< PixRGB< byte > > auxImage, bool fast=false)
	Put image one at a time into tracker and get back tracker image output.
void	SIMtrackImageMulti (std::vector< Image< PixRGB< byte > > > *image, int instances)
	put image set into tracker and let it sort all this stuff out
void	SIMmergeImages (Image< PixRGB< byte > > *image)
	merge qualities of images if needed
void	SIMupdateVergance (float distance, float gaussBase=36)
	update the vergance distance if needed for cameras
void	SIMverganceSpring (int instance, int current, bool doTracked)
	apply probabalistic model over multiple stereo cameras for vergance
void	SIMgetImageTrackXY (int x, int y, int instance)
	get back X and Y coords for tracking
void	SIMgetImageTrackXY2 (int x, int y, int instance)
	get back X and Y coords for tracking
void	SIMgetImageTrackXYMerge (int x, int y)
	get back merged X and Y coords for tracking
bool	SIMreturnLOT (int instance)
	return true if loss of track
float	SIMreturnCameraProb (int instance)
	return the P of a good track
bool	SIMdoMoveCamera (int instance, float doPan, float doTilt)
	return if camera should move based upon maximum LOT's
Image< byte >	SIMreturnCandidateImage (int instance)
	return the image of candidate pixels
void	SIMSetCluster (int sizeX, int sizeY, int instances, float hweight, float sweight, float vweight)
	Set up color clustering for tracker.
void	SIMclusterColor (Image< PixRGB< float > > image, int instance)
	find object in scene that's a best match using clustering
bool	SIMstereoMatch (PixelPoint points[2], CameraParams params[2], Point3D *retPoint)
	Project corresponding points in 3D.
double	SIMPgauss (double X, double Xbar, double std)
	return the probability of X given Xbar and std using gaussian

Detailed Description

Definition at line 54 of file segmentImageMerge2.H.

Constructor & Destructor Documentation

segmentImageMerge2::segmentImageMerge2 ( int instances )

default constructor, also sets up number of tracker instances

set the weights in cluster for Hue

set the weights in cluster for Sat

set the weights in cluster for Val

set initial standard deviation for color tracking after Hue

set initial standard deviation for color tracking after Sat

set initial standard deviation for color tracking after Val

Definition at line 457 of file segmentImageMerge2.C.

References blobProp::BP_bound, blobProp::BP_clusterColorStdH, blobProp::BP_clusterColorStdS, blobProp::BP_clusterColorStdV, blobProp::BP_clusterWeightH, blobProp::BP_clusterWeightS, blobProp::BP_clusterWeightSpat, blobProp::BP_clusterWeightV, blobProp::BP_LOTbound, blobProp::BP_maxFrameSize, blobProp::BP_maxMass, blobProp::BP_minClusterSize, blobProp::BP_minMass, blobProp::BP_ratioMax, blobProp::BP_ratioMin, blobProp::BP_softBound, blobProp::BP_softRatioMax, blobProp::BP_softRatioMin, CAMERAMU1, ERRINTERVAL, readConfig::getItemValueF(), and readConfig::openFile().

Member Function Documentation

void segmentImageMerge2::resetAll ( int instances )

set the weights in cluster for Hue

set the weights in cluster for Sat

set the weights in cluster for Val

set initial standard deviation for color tracking after Hue

set initial standard deviation for color tracking after Sat

set initial standard deviation for color tracking after Val

Definition at line 333 of file segmentImageMerge2.C.

References blobProp::BP_bound, blobProp::BP_clusterColorStdH, blobProp::BP_clusterColorStdS, blobProp::BP_clusterColorStdV, blobProp::BP_clusterWeightH, blobProp::BP_clusterWeightS, blobProp::BP_clusterWeightSpat, blobProp::BP_clusterWeightV, blobProp::BP_LOTbound, blobProp::BP_maxFrameSize, blobProp::BP_maxMass, blobProp::BP_minClusterSize, blobProp::BP_minMass, blobProp::BP_ratioMax, blobProp::BP_ratioMin, blobProp::BP_softBound, blobProp::BP_softRatioMax, blobProp::BP_softRatioMin, CAMERAMU1, ERRINTERVAL, readConfig::getItemValueF(), and readConfig::openFile().

void segmentImageMerge2::SIMclusterColor	(	Image< PixRGB< float > >	image,
		int	instance
	)

find object in scene that's a best match using clustering

run this to find out which color to use for tracking.This will take and image and cluster it looking for the best color match in the image to track

Definition at line 1037 of file segmentImageMerge2.C.

References ASSERT, Image< T >::beginw(), blobProp::BP_clusterColorStdH, blobProp::BP_clusterColorStdS, blobProp::BP_clusterColorStdV, blobProp::BP_clusterWeightH, blobProp::BP_clusterWeightS, blobProp::BP_clusterWeightSpat, blobProp::BP_clusterWeightV, blobProp::BP_minClusterSize, decXY(), Image< T >::endw(), Image< T >::getWidth(), image, newImage, NPclassify2< FLOAT >::NPaddSpace(), NPclassify2< FLOAT >::NPclassifySpaceNew(), NPclassify2< FLOAT >::NPgetClass(), NPclassify2< FLOAT >::NPgetClassSize(), NPclassify2< FLOAT >::NPgetFeature(), NPclassify2< FLOAT >::NPgetStemNumber(), NPclassify2< FLOAT >::NPresetSpace(), SIMPgauss(), segmentImage2::SIsetHue(), segmentImage2::SIsetSat(), and segmentImage2::SIsetVal().

bool segmentImageMerge2::SIMdoMoveCamera	(	int	instance,
		float *	doPan,
		float *	doTilt
	)

return if camera should move based upon maximum LOT's

Definition at line 996 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMgetImageTrackXY	(	int *	x,
		int *	y,
		int	instance
	)

get back X and Y coords for tracking

Definition at line 968 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMgetImageTrackXY2	(	int *	x,
		int *	y,
		int	instance
	)

get back X and Y coords for tracking

Definition at line 974 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMgetImageTrackXYMerge	(	int *	x,
		int *	y
	)

get back merged X and Y coords for tracking

Definition at line 980 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMmergeImages ( Image< PixRGB< byte > > * image )

merge qualities of images if needed

Definition at line 791 of file segmentImageMerge2.C.

References drawCircle(), and image.

double segmentImageMerge2::SIMPgauss	(	double	X,
		double	Xbar,
		double	std
	)

return the probability of X given Xbar and std using gaussian

Definition at line 1279 of file segmentImageMerge2.C.

References exp(), and sqrt().

Referenced by SIMclusterColor().

float segmentImageMerge2::SIMreturnCameraProb ( int instance )

return the P of a good track

Definition at line 991 of file segmentImageMerge2.C.

Image< byte > segmentImageMerge2::SIMreturnCandidateImage ( int instance )

return the image of candidate pixels

Definition at line 1006 of file segmentImageMerge2.C.

References segmentImage2::SIreturnNormalizedCandidates().

bool segmentImageMerge2::SIMreturnLOT ( int instance )

return true if loss of track

Definition at line 986 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMsetAdapt	(	float	ha,
		bool	haSTD,
		float	sa,
		bool	saSTD,
		float	va,
		bool	vaSTD,
		int	instance,
		bool	useCluster
	)

set initial adaptive color thresholds for tracking

Definition at line 617 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMsetAdaptBound	(	float	Hupper,
		float	Hlower,
		float	Supper,
		float	Slower,
		float	Vupper,
		float	Vlower,
		int	instance
	)

set up hard bounds for color adaptation

Parameters:

	Hupper	upper bound for adaptive hue
	Hlower	lower bound for adaptive hue
	Supper	upper bound for adaptive saturation
	Slower	lower bound for adaptive saturation
	Vupper	upper bound for adaptive intensity
	Vlower	lower bound for adaptive intensity
	instance	Which tracker to apply these settings to

Definition at line 631 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMsetBoxColor	(	int	r,
		int	g,
		int	b,
		int	instance
	)

set the tracking colot for the bounding box set the tracking color for mass circle

Parameters:

	r	this is the red color for the output bounding box
	g	this is the green color for the output bounding box
	b	this is the blue color for the output bounding box
	instance	this is the tracker to apply these settings to

Definition at line 590 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMsetCameraPosition	(	float	pan,
		float	tilt,
		int	instance,
		bool	stats = `false`
	)

tell this method the pan and tilt postition the camera is moving to

Definition at line 644 of file segmentImageMerge2.C.

References ERRINTERVAL, and sqrt().

void segmentImageMerge2::SIMsetCircleColor	(	int	r,
		int	g,
		int	b,
		int	instance
	)

set the tracking color for mass circle

Parameters:

	r	this is the red color for the output circle
	g	this is the green color for the output circle
	b	this is the blue color for the output circle
	instance	this is the tracker to apply these settings to

Definition at line 582 of file segmentImageMerge2.C.

void segmentImageMerge2::SIMSetCluster	(	int	sizeX,
		int	sizeY,
		int	instances,
		float	hweight,
		float	sweight,
		float	vweight
	)

Set up color clustering for tracker.

Call this method before calling SIMclsuterColor to set it up. Do this only once.

Parameters:

	sizeX	This is the width of the image to be processed
	sizeY	this is the height of the image to be processed
	instances	How many tracker instances are being used
	hweight	this is the weight to be given to hue in analysis
	sweight	this is the weight to be given to saturation in analysis
	vweight	this is the weight to be given to luminance in analysis

Definition at line 1011 of file segmentImageMerge2.C.

References NPclassify2< FLOAT >::NPresizeSpace(), NPclassify2< FLOAT >::NPsetup(), and readConfig::openFile().

void segmentImageMerge2::SIMsetFrame	(	int	x1,
		int	y1,
		int	x2,
		int	y2,
		int	realX,
		int	realY,
		int	instance
	)

set frame size for color tracker

Definition at line 695 of file segmentImageMerge2.C.

References segmentImage2::SIsetFrame().

void segmentImageMerge2::SIMsetTrackColor	(	float	h,
		float	hstd,
		float	s,
		float	sstd,
		float	v,
		float	vstd,
		int	instance,
		bool	adapt,
		int	avg
	)

set default tracking color values

Parameters:

	h	initial hue value
	hstd	initial boundary in standard deviation for h
	s	initial satuaration value
	sstd	initial boundary in standard deviation for s
	v	initial intensity value
	vstd	initial boundary in standard deviation for v
	instance	which tracker instance to apply these settings too
	adapt	tells if you want these values to adapt

Definition at line 598 of file segmentImageMerge2.C.

References segmentImage2::SIsetHSVavg(), segmentImage2::SIsetHue(), segmentImage2::SIsetSat(), and segmentImage2::SIsetVal().

bool segmentImageMerge2::SIMstereoMatch	(	PixelPoint	points[2],
		CameraParams	params[2],
		Point3D *	retPoint
	)

Project corresponding points in 3D.

Definition at line 1208 of file segmentImageMerge2.C.

References Point3D< T >::x.

void segmentImageMerge2::SIMtrackImage	(	Image< PixRGB< byte > >	input,
		Image< PixRGB< byte > > *	image,
		int	instance,
		Image< PixRGB< byte > > *	auxImage,
		bool	fast = `false`
	)

Put image one at a time into tracker and get back tracker image output.

Parameters:

	input	this is the raw input image
	image	this is a pointer to the output image
	outputI	this is a pointer the the candidate pixel image
	instance	This is the image instance to which you are using
	auxImage	this is a pointer to the HSV threshold bar image

Definition at line 702 of file segmentImageMerge2.C.

References ASSERT, decXY(), image, segmentImage2::SIcalcMassCenter(), and segmentImage2::SIsegment().

void segmentImageMerge2::SIMtrackImageMulti	(	std::vector< Image< PixRGB< byte > > > *	image,
		int	instances
	)

put image set into tracker and let it sort all this stuff out

this method will track over multiple stereo pair linked cameras and produce blobs to track that agree with assumptions of stereo tracking for instance vergance and spacial locality.

Parameters:

	input	this is the raw input image
	image	this is a pointer to the output image
	outputI	this is a pointer the the candidate pixel image
	instance	This is the image instance to which you are using
	auxImage	this is a pointer to the HSV threshold bar image

Definition at line 739 of file segmentImageMerge2.C.

References ASSERT, decXY(), image, segmentImage2::SIcalcMassCenter(), SIMupdateVergance(), SIMverganceSpring(), and segmentImage2::SIsegment().

void segmentImageMerge2::SIMupdateVergance	(	float	distance,
		float	gaussBase = `36`
	)

update the vergance distance if needed for cameras

feed in the new distance to target and extact P over expected vergance

Parameters:

distance

is a measure in inches (sorry)

Definition at line 816 of file segmentImageMerge2.C.

References DISTANCE.

Referenced by SIMtrackImageMulti().

void segmentImageMerge2::SIMverganceSpring	(	int	instance,
		int	current,
		bool	doTracked
	)

apply probabalistic model over multiple stereo cameras for vergance

This method will calculate the most likely blob from a set of blobs based upon the vergance of the cameras. That is, joined stereo pairs should. experiance vergance to a target. This method creates virtual springs to pull the cameras into vergance with each other based upon the probability that a target will bring a camera into vergance with the other cameras. To use this method you must have defined a probabalistic model over vergance in the cameras.

Parameters:

	instance	this is the instance of the track, in this case the camera
	doTracked	this is true if you wish to apply to cameras that \ are not currently in loss of track (LOT)

Definition at line 834 of file segmentImageMerge2.C.

References abs(), cameraConf::fieldAdjustmentX, cameraConf::fieldAdjustmentY, stats< T >::gauss(), LOTMAX, segmentImage2::SIgetCenterX(), segmentImage2::SInumberBlobs(), cameraConf::Xfield, cameraConf::Xpixel, cameraConf::Yfield, and cameraConf::Ypixel.

Referenced by SIMtrackImageMulti().

The documentation for this class was generated from the following files:

src/VFAT/segmentImageMerge2.H
src/VFAT/segmentImageMerge2.C

segmentImageMerge2 Class Reference

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation