jafar::slam::ImagePluckerFeatureObserveModel Class Reference

Sizeless 3D segment using the homogenous coordinates parametrisation observed in a perspective image. More...

---

Detailed Description

Sizeless 3D segment using the homogenous coordinates parametrisation observed in a perspective image.

Class ImagePluckerFeatureObserveModel

Definition at line 448 of file segmentFeature.hpp.

#include <segmentFeature.hpp>

Inheritance diagram for jafar::slam::ImagePluckerFeatureObserveModel:

List of all members.

Public Member Functions
	ImagePluckerFeatureObserveModel (SegmentFeatureModel &featureModel_, jafar::camera::CameraPinhole const &camera_)
Observation::ObservationType	typeObs () const
void	setPixCov (jblas::vec const &pixCov_, double stabilizingFactor=1.0)
void	initInternalState (BaseFeature &feature, jafar::geom::T3DEuler const &robotPose_, Observation const &obs_) const
	Init extremities of the segment.
void	updateInternalState (BaseFeature &feature, jafar::geom::T3DEuler const &robotPose_, Observation const &obs_) const
	Update extremities of the segment.
void	predictExtObs (jblas::vec_range const &pose, jblas::sym_mat_range const &poseCov, jblas::vec3 const &ext, jblas::sym_mat const &extCov, jblas::vec &zPredExt, jblas::sym_mat &zPredExtCov)
	Predict the extremes of the observed line, with their covariance in the image plane.
void	predictExtObsJac (jblas::vec_range const &pose, jblas::vec3 const Ext, jblas::vec &eInImage, jblas::mat &JeR, jblas::mat &JeExt)
	Predict the extreme and Jacobians.
void	postInitCovariance (jblas::sym_mat_range &P, jblas::mat const &Jx)
	Assign covariances of the unobserved part.
void	setup (double beta0=0, double beta1=0, double sigma_beta=0.5)
	Set up initial parameters.
Protected Member Functions
virtual jblas::vec const &	homogeneousToImageModel (jblas::vec3 const &l_)=0
	homogeneous to Image 2D line representation conversion
virtual void	homogeneousToImageModelJac (jblas::vec3 const &l_, jblas::mat &Jl) const =0
	homogeneous to Image 2D line representation -- Jacobians
virtual jblas::vec3	imageModelToHomogeneous (jblas::vec const &rt_)=0
	Image representation to homogeneous transform.
virtual void	imageModelToHomogeneousJac (jblas::vec const &rt_, jblas::mat &HMrt)=0
	Rho-theta to homogeneous transform -- Jacobian.
virtual void	imageModelToHomogeneousJac (jblas::vec const &rt_, jblas::vec3 &l, jblas::mat &HMrt)=0
	Rho-theta to homogeneous transform -- Jacobian.
jblas::vec3	planeBaseToDirVector (jblas::vec3 const &an_, jblas::vec2 const &beta_)
	Plane sub-vector a to plane base vector beta.
void	planeBaseToDirVectorJac (jblas::vec_range const &an_, jblas::vec2 const &beta_, jblas::vec_range &b, jblas::mat33 &B_a)
	Plane sub-vector a to plane base vector -- Jacobians.
boost::tuple< double, double >	computeExtremitiesAbscissa (SegmentFeature const &segFeature_, geom::T3DEuler const &robotPose_, SegmentObservation const &obs_) const
void	segmentExtremitiesPluckerLines (SegmentObservation const &segObs_, jblas::vec &lineExt1Sensor, jblas::vec &lineExt2Sensor) const
Static Protected Member Functions
static boost::tuple< double, double, double >	pluckerLinesDistance (jblas::vec const &line1_, jblas::vec const &line2_)
Protected Attributes
SegmentFeatureModel &	segmentFeatureModel
	Feature model.
jafar::camera::CameraPinhole	camera
	Perspective camera.
jblas::mat33	Pl
	Plucker camera matrix.
jblas::mat33	Plinv
	Plucker camera inverse matrix.
jblas::vec2	beta
	initial plane base vector
jblas::mat22	covar_beta
	covariances matrix of initial plane base vector
jblas::mat	JbBeta
	Jacobian of b wrt beta.
jblas::sym_mat	p_pixCov
	covariance on pixel coordinate (u,v)

---

Member Function Documentation

virtual jblas::vec3 jafar::slam::ImagePluckerFeatureObserveModel::imageModelToHomogeneous

(

jblas::vec const &

rt_

)

[protected, pure virtual]

Image representation to homogeneous transform.

Parameters:

rt_	input 2D line representation -- e.g. rt = (rho,theta)'

Returns:

The homogeneous line in P3 -- l = (a,b,c)'

Implemented in jafar::slam::ScaledHomogeneousImagePluckerFeatureObserveModel, and jafar::slam::RhoThetaImagePluckerFeatureObserveModel.

virtual void jafar::slam::ImagePluckerFeatureObserveModel::imageModelToHomogeneousJac	(	jblas::vec const &	rt_,
		jblas::mat &	HMrt
	)		[protected, pure virtual]

Rho-theta to homogeneous transform -- Jacobian.

Parameters:

rt_	input 2D line representation
HMrt	output transformation Jacobian

Implemented in jafar::slam::ScaledHomogeneousImagePluckerFeatureObserveModel, and jafar::slam::RhoThetaImagePluckerFeatureObserveModel.

virtual void jafar::slam::ImagePluckerFeatureObserveModel::imageModelToHomogeneousJac	(	jblas::vec const &	rt_,
		jblas::vec3 &	l,
		jblas::mat &	HMrt
	)		[protected, pure virtual]

Rho-theta to homogeneous transform -- Jacobian.

Parameters:

rt_	input 2D line representation
l	output homogeneous line representation
HMrt	output transformation Jacobian

Implemented in jafar::slam::ScaledHomogeneousImagePluckerFeatureObserveModel, and jafar::slam::RhoThetaImagePluckerFeatureObserveModel.

void jafar::slam::ImagePluckerFeatureObserveModel::postInitCovariance	(	jblas::sym_mat_range &	P,
		jblas::mat const &	Jx
	)		[virtual]

Assign covariances of the unobserved part.

Parameters:

P	covariance of the landmark to initialize in global frame
Jx	Jacobian of the feature in global frame wrt the feat. in robot frame

Reimplemented from jafar::slam::FeatureObserveModel.

void jafar::slam::ImagePluckerFeatureObserveModel::predictExtObs	(	jblas::vec_range const &	pose,
		jblas::sym_mat_range const &	poseCov,
		jblas::vec3 const &	ext,
		jblas::sym_mat const &	extCov,
		jblas::vec &	zPredExt,
		jblas::sym_mat &	zPredExtCov
	)

Predict the extremes of the observed line, with their covariance in the image plane.

Parameters:

pose	the robot pose
poseCov	the robot pose covariance matrix
ext	one of the extremities in 3D space
extCov	the covariances matrix of the extreme in 3D space
zPredExt	the predicted extreme in the 2D image
zPredExtCov	the covariances matrix of the predicted extreme CAUTION: THIS IMPLEMENTATION IGNORES THE CROSS-VARIANCES ROBOT/FEATURE AND LEADS TO TOO BIG ELLIPSES IN THE IMAGE.

void jafar::slam::ImagePluckerFeatureObserveModel::predictExtObsJac	(	jblas::vec_range const &	pose,
		jblas::vec3 const	Ext,
		jblas::vec &	eInImage,
		jblas::mat &	JeR,
		jblas::mat &	JeExt
	)

Predict the extreme and Jacobians.

Parameters:

pose	the robot pose
Ext	one of the extremities in 3D space
eInImage	the predicted extreme in the 2D image
JeR	and
JeExt	Jacobians of the prediction with respect to pose and extremity

void jafar::slam::ImagePluckerFeatureObserveModel::setup	(	double	beta0 = 0,
		double	beta1 = 0,
		double	sigma_beta = 0.5
	)

Set up initial parameters.

Parameters:

beta0	The first component of the lines director vector for initialization
beta1	the second component
sigma_beta	The standard deviation of beta0 and beta1

---

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

• modules/slam/include/slam/segmentFeature.hpp