vpgl_bundle_adjust_lsqr Class Reference

base class bundle adjustment sparse least squares function. More...

#include <vpgl_bundle_adjust_lsqr.h>

Inheritance diagram for vpgl_bundle_adjust_lsqr:

List of all members.

Public Types
enum	UseGradient
enum	UseWeights
Public Member Functions
	vpgl_bundle_adjust_lsqr (unsigned int num_params_per_a, unsigned int num_params_per_b, unsigned int num_params_c, const vcl_vector< vgl_point_2d< double > > &image_points, const vcl_vector< vcl_vector< bool > > &mask)
	Constructor.
	vpgl_bundle_adjust_lsqr (unsigned int num_params_per_a, unsigned int num_params_per_b, unsigned int num_params_c, const vcl_vector< vgl_point_2d< double > > &image_points, const vcl_vector< vnl_matrix< double > > &inv_covars, const vcl_vector< vcl_vector< bool > > &mask)
	Constructor.
virtual	~vpgl_bundle_adjust_lsqr ()
virtual void	f (vnl_vector< double > const &a, vnl_vector< double > const &b, vnl_vector< double > const &c, vnl_vector< double > &e)
	Compute all the reprojection errors.
virtual void	fij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_vector< double > &fij)
	Compute the residuals from the ith component of a, the jth component of b, and all of c.
virtual void	jac_blocks (vnl_vector< double > const &a, vnl_vector< double > const &b, vnl_vector< double > const &c, vcl_vector< vnl_matrix< double > > &A, vcl_vector< vnl_matrix< double > > &B, vcl_vector< vnl_matrix< double > > &C)
	Compute the sparse Jacobian in block form.
virtual void	jac_Aij (unsigned int i, unsigned int j, vnl_double_3x4 const &Pi, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Aij)=0
	compute the Jacobian Aij.
virtual void	jac_Bij (unsigned int i, unsigned int j, vnl_double_3x4 const &Pi, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Bij)=0
	compute the Jacobian Bij.
virtual void	jac_Cij (unsigned int i, unsigned int j, vnl_double_3x4 const &Pi, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Cij)=0
	compute the Jacobian Cij.
void	compute_weight_ij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_vector< double > const &fij, double &weight)
	Use an M-estimator to compute weights.
void	set_residual_scale (double scale)
	set the residual scale for the robust estimation.
vgl_homg_point_3d< double >	param_to_point (int j, const vnl_vector< double > &b, const vnl_vector< double > &c) const
	construct the
virtual vgl_homg_point_3d< double >	param_to_point (int j, const double *bj, const vnl_vector< double > &c) const =0
	construct the
vnl_vector_fixed< double, 4 >	param_to_pt_vector (int j, const vnl_vector< double > &b, const vnl_vector< double > &c) const
	construct the
virtual vnl_vector_fixed < double, 4 >	param_to_pt_vector (int j, const double *bj, const vnl_vector< double > &c) const =0
	construct the
vpgl_perspective_camera< double >	param_to_cam (int i, const vnl_vector< double > &a, const vnl_vector< double > &c) const
	construct the
virtual vpgl_perspective_camera < double >	param_to_cam (int i, const double *ai, const vnl_vector< double > &c) const =0
	construct the
vnl_double_3x4	param_to_cam_matrix (int i, const vnl_vector< double > &a, const vnl_vector< double > &c) const
	construct the
virtual vnl_double_3x4	param_to_cam_matrix (int i, const double *ai, const vnl_vector< double > &c) const =0
	compute the 3x4 matrix of camera
void	reset ()
	reset the weights.
void	throw_failure ()
void	clear_failure ()
virtual void	fd_jac_blocks (vnl_vector< double > const &a, vnl_vector< double > const &b, vnl_vector< double > const &c, vcl_vector< vnl_matrix< double > > &A, vcl_vector< vnl_matrix< double > > &B, vcl_vector< vnl_matrix< double > > &C, double stepsize)
virtual void	compute_weights (vnl_vector< double > const &a, vnl_vector< double > const &b, vnl_vector< double > const &c, vnl_vector< double > const &f, vnl_vector< double > &weights)
virtual void	apply_weights (vnl_vector< double > const &weights, vnl_vector< double > &f)
virtual void	apply_weights (vnl_vector< double > const &weights, vcl_vector< vnl_matrix< double > > &A, vcl_vector< vnl_matrix< double > > &B, vcl_vector< vnl_matrix< double > > &C)
virtual void	jac_Aij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Aij)
virtual void	jac_Bij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Bij)
virtual void	jac_Cij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Cij)
void	fd_jac_Aij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Aij, double stepsize)
void	fd_jac_Bij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Bij, double stepsize)
void	fd_jac_Cij (int i, int j, vnl_vector< double > const &ai, vnl_vector< double > const &bj, vnl_vector< double > const &c, vnl_matrix< double > &Cij, double stepsize)
virtual void	apply_weight_ij (int i, int j, double const &weight, vnl_vector< double > &fij)
virtual void	apply_weight_ij (int i, int j, double const &weight, vnl_matrix< double > &Aij, vnl_matrix< double > &Bij, vnl_matrix< double > &Cij)
virtual void	trace (int iteration, vnl_vector< double > const &a, vnl_vector< double > const &b, vnl_vector< double > const &c, vnl_vector< double > const &e)
unsigned int	number_of_params_a (int i) const
unsigned int	number_of_params_b (int j) const
unsigned int	number_of_params_c () const
unsigned int	number_of_residuals (int k) const
unsigned int	number_of_residuals (int i, int j) const
unsigned int	index_a (int i) const
unsigned int	index_b (int j) const
unsigned int	index_e (int k) const
unsigned int	number_of_a () const
unsigned int	number_of_b () const
unsigned int	number_of_e () const
bool	has_gradient () const
bool	has_weights () const
const vnl_crs_index &	residual_indices () const
Static Public Member Functions
static vnl_double_3x3	rod_to_matrix (vnl_vector< double > const &r)
	Fast conversion of rotation from Rodrigues vector to matrix.
static void	jac_inhomg_3d_point (vnl_double_3x4 const &P, vnl_vector< double > const &pt, vnl_matrix< double > &J)
	compute the 2x3 Jacobian of camera projection with respect to point location df/dpt where $f(pt) = P*pt$.
static void	jac_camera_center (vnl_double_3x3 const &M, vnl_vector< double > const &C, vnl_vector< double > const &pt, vnl_matrix< double > &J)
	compute the 2x3 Jacobian of camera projection with respect to camera center df/dC where $f(C) = [M | -M*C]*pt$.
static void	jac_camera_rotation (vnl_double_3x3 const &K, vnl_vector< double > const &C, vnl_vector< double > const &r, vnl_vector< double > const &pt, vnl_matrix< double > &J)
	compute the 2x3 Jacobian of camera projection with respect to camera rotation df/dr where $f(r) = K*rod_to_matrix(r)*[I | -C]*pt$.
Public Attributes
	no_gradient
	use_gradient
	no_weights
	use_weights
bool	failure
Protected Attributes
vcl_vector< vgl_point_2d < double > >	image_points_
	The corresponding points in the image.
vcl_vector< vnl_matrix< double > >	factored_inv_covars_
	The Cholesky factored inverse covariances for each image point.
bool	use_covars_
	Flag to enable covariance weighted errors.
double	scale2_
	The square of the scale of the robust estimator.
int	iteration_count_
vnl_crs_index	residual_indices_
vcl_vector< unsigned int >	indices_a_
vcl_vector< unsigned int >	indices_b_
unsigned int	num_params_c_
vcl_vector< unsigned int >	indices_e_
bool	use_gradient_
bool	use_weights_

---

Detailed Description

base class bundle adjustment sparse least squares function.

Definition at line 27 of file vpgl_bundle_adjust_lsqr.h.

---

Constructor & Destructor Documentation

vpgl_bundle_adjust_lsqr::vpgl_bundle_adjust_lsqr	(	unsigned int	num_params_per_a,
		unsigned int	num_params_per_b,
		unsigned int	num_params_c,
		const vcl_vector< vgl_point_2d< double > > &	image_points,
		const vcl_vector< vcl_vector< bool > > &	mask
	)

Constructor.

Note:

image points are not homogeneous because they require finite points to measure projection error

Definition at line 16 of file vpgl_bundle_adjust_lsqr.cxx.

vpgl_bundle_adjust_lsqr::vpgl_bundle_adjust_lsqr	(	unsigned int	num_params_per_a,
		unsigned int	num_params_per_b,
		unsigned int	num_params_c,
		const vcl_vector< vgl_point_2d< double > > &	image_points,
		const vcl_vector< vnl_matrix< double > > &	inv_covars,
		const vcl_vector< vcl_vector< bool > > &	mask
	)

Constructor.

Each image point is assigned an inverse covariance (error projector) matrix

Note:

image points are not homogeneous because they require finite points to measure projection error

Definition at line 37 of file vpgl_bundle_adjust_lsqr.cxx.

virtual vpgl_bundle_adjust_lsqr::~vpgl_bundle_adjust_lsqr

(

)

[inline, virtual]

Definition at line 51 of file vpgl_bundle_adjust_lsqr.h.

---

Member Function Documentation

void vpgl_bundle_adjust_lsqr::compute_weight_ij	(	int	i,
		int	j,
		vnl_vector< double > const &	ai,
		vnl_vector< double > const &	bj,
		vnl_vector< double > const &	c,
		vnl_vector< double > const &	fij,
		double &	weight
	)		[virtual]

Use an M-estimator to compute weights.

Reimplemented from vnl_sparse_lst_sqr_function.

Definition at line 198 of file vpgl_bundle_adjust_lsqr.cxx.

void vpgl_bundle_adjust_lsqr::f	(	vnl_vector< double > const &	a,
		vnl_vector< double > const &	b,
		vnl_vector< double > const &	c,
		vnl_vector< double > &	e
	)		[virtual]

Compute all the reprojection errors.

Given the parameter vectors a, b, and c, compute the vector of residuals e. e has been sized appropriately before the call. The parameters in for each camera are in a. The 3D point parameters are in b. Global parameters are in c.

Given the parameter vectors a, b, and c, compute the vector of residuals e. e has been sized appropriately before the call. The parameters in a for each camera are {wx, wy, wz, tx, ty, tz} where w is the Rodrigues vector of the rotation and t is the translation. The parameters in b for each 3D point are {px, py, pz} the non-homogeneous position.

Construct the ith camera.

Reimplemented from vnl_sparse_lst_sqr_function.

Definition at line 85 of file vpgl_bundle_adjust_lsqr.cxx.

void vpgl_bundle_adjust_lsqr::fij	(	int	i,
		int	j,
		vnl_vector< double > const &	ai,
		vnl_vector< double > const &	bj,
		vnl_vector< double > const &	c,
		vnl_vector< double > &	fij
	)		[virtual]

Compute the residuals from the ith component of a, the jth component of b, and all of c.

Compute the residuals from the ith component of a and the jth component of b.

This function is not normally used because f() has a self-contained efficient implementation. It is used for finite-differencing if the gradient is marked as unavailable

This is not normally used because f() has a self-contained efficient implementation It is used for finite-differencing if the gradient is marked as unavailable

Construct the ith camera.

Reimplemented from vnl_sparse_lst_sqr_function.

Definition at line 127 of file vpgl_bundle_adjust_lsqr.cxx.

virtual void vpgl_bundle_adjust_lsqr::jac_Aij	(	unsigned int	i,
		unsigned int	j,
		vnl_double_3x4 const &	Pi,
		vnl_vector< double > const &	ai,
		vnl_vector< double > const &	bj,
		vnl_vector< double > const &	c,
		vnl_matrix< double > &	Aij
	)		[pure virtual]

compute the Jacobian Aij.

Implemented in vpgl_ba_fixed_k_lsqr.

virtual void vpgl_bundle_adjust_lsqr::jac_Bij	(	unsigned int	i,
		unsigned int	j,
		vnl_double_3x4 const &	Pi,
		vnl_vector< double > const &	ai,
		vnl_vector< double > const &	bj,
		vnl_vector< double > const &	c,
		vnl_matrix< double > &	Bij
	)		[pure virtual]

compute the Jacobian Bij.

Implemented in vpgl_ba_fixed_k_lsqr.

void vpgl_bundle_adjust_lsqr::jac_blocks	(	vnl_vector< double > const &	a,
		vnl_vector< double > const &	b,
		vnl_vector< double > const &	c,
		vcl_vector< vnl_matrix< double > > &	A,
		vcl_vector< vnl_matrix< double > > &	B,
		vcl_vector< vnl_matrix< double > > &	C
	)		[virtual]

Compute the sparse Jacobian in block form.

Construct the ith camera.

Reimplemented from vnl_sparse_lst_sqr_function.

Definition at line 157 of file vpgl_bundle_adjust_lsqr.cxx.

void vpgl_bundle_adjust_lsqr::jac_camera_center	(	vnl_double_3x3 const &	M,
		vnl_vector< double > const &	C,
		vnl_vector< double > const &	pt,
		vnl_matrix< double > &	J
	)		[static]

compute the 2x3 Jacobian of camera projection with respect to camera center df/dC where $f(C) = [M | -M*C]*pt$.

Definition at line 249 of file vpgl_bundle_adjust_lsqr.cxx.

void vpgl_bundle_adjust_lsqr::jac_camera_rotation	(	vnl_double_3x3 const &	K,
		vnl_vector< double > const &	C,
		vnl_vector< double > const &	r,
		vnl_vector< double > const &	pt,
		vnl_matrix< double > &	J
	)		[static]

compute the 2x3 Jacobian of camera projection with respect to camera rotation df/dr where $f(r) = K*rod_to_matrix(r)*[I | -C]*pt$.

Here r is a Rodrigues vector, K is an upper triangular calibration matrix

Definition at line 265 of file vpgl_bundle_adjust_lsqr.cxx.

virtual void vpgl_bundle_adjust_lsqr::jac_Cij	(	unsigned int	i,
		unsigned int	j,
		vnl_double_3x4 const &	Pi,
		vnl_vector< double > const &	ai,
		vnl_vector< double > const &	bj,
		vnl_vector< double > const &	c,
		vnl_matrix< double > &	Cij
	)		[pure virtual]

compute the Jacobian Cij.

Implemented in vpgl_ba_fixed_k_lsqr.

void vpgl_bundle_adjust_lsqr::jac_inhomg_3d_point	(	vnl_double_3x4 const &	P,
		vnl_vector< double > const &	pt,
		vnl_matrix< double > &	J
	)		[static]

compute the 2x3 Jacobian of camera projection with respect to point location df/dpt where $f(pt) = P*pt$.

Definition at line 217 of file vpgl_bundle_adjust_lsqr.cxx.

vpgl_perspective_camera<double> vpgl_bundle_adjust_lsqr::param_to_cam	(	int	i,
		const vnl_vector< double > &	a,
		const vnl_vector< double > &	c
	)		const [inline]

construct the

Parameters:

i-th	perspective camera from parameter vector
a.

Definition at line 153 of file vpgl_bundle_adjust_lsqr.h.

virtual vpgl_perspective_camera<double> vpgl_bundle_adjust_lsqr::param_to_cam	(	int	i,
		const double *	ai,
		const vnl_vector< double > &	c
	)		const [pure virtual]

construct the

Parameters:

i-th	perspective camera from a pointer to the i-th parameter of
ai	and parameters
c.

Implemented in vpgl_ba_fixed_k_lsqr.

vnl_double_3x4 vpgl_bundle_adjust_lsqr::param_to_cam_matrix	(	int	i,
		const vnl_vector< double > &	a,
		const vnl_vector< double > &	c
	)		const [inline]

construct the

Parameters:

i-th	perspective camera matrix from parameter vectors
a	and
c.

Definition at line 167 of file vpgl_bundle_adjust_lsqr.h.

virtual vnl_double_3x4 vpgl_bundle_adjust_lsqr::param_to_cam_matrix	(	int	i,
		const double *	ai,
		const vnl_vector< double > &	c
	)		const [pure virtual]

compute the 3x4 matrix of camera

Parameters:

i	from a pointer to the i-th parameter of
ai	and parameters
c.

Implemented in vpgl_ba_fixed_k_lsqr.

vgl_homg_point_3d<double> vpgl_bundle_adjust_lsqr::param_to_point	(	int	j,
		const vnl_vector< double > &	b,
		const vnl_vector< double > &	c
	)		const [inline]

construct the

Parameters:

j-th	3D point from parameter vector
b	and
c.

Definition at line 123 of file vpgl_bundle_adjust_lsqr.h.

virtual vgl_homg_point_3d<double> vpgl_bundle_adjust_lsqr::param_to_point	(	int	j,
		const double *	bj,
		const vnl_vector< double > &	c
	)		const [pure virtual]

construct the

Parameters:

j-th	perspective camera from a pointer to the j-th parameter of
bj	and parameters
c.

Implemented in vpgl_ba_fixed_k_lsqr.

vnl_vector_fixed<double,4> vpgl_bundle_adjust_lsqr::param_to_pt_vector	(	int	j,
		const vnl_vector< double > &	b,
		const vnl_vector< double > &	c
	)		const [inline]

construct the

Parameters:

j-th	3D point from parameter vector
b	and
c.

Definition at line 138 of file vpgl_bundle_adjust_lsqr.h.

virtual vnl_vector_fixed<double,4> vpgl_bundle_adjust_lsqr::param_to_pt_vector	(	int	j,
		const double *	bj,
		const vnl_vector< double > &	c
	)		const [pure virtual]

construct the

Parameters:

j-th	perspective camera from a pointer to the j-th parameter of
bj	and parameters
c.

Implemented in vpgl_ba_fixed_k_lsqr.

void vpgl_bundle_adjust_lsqr::reset

(

)

[inline]

reset the weights.

Definition at line 180 of file vpgl_bundle_adjust_lsqr.h.

vnl_double_3x3 vpgl_bundle_adjust_lsqr::rod_to_matrix

(

vnl_vector< double > const &

r

)

[static]

Fast conversion of rotation from Rodrigues vector to matrix.

Definition at line 359 of file vpgl_bundle_adjust_lsqr.cxx.

void vpgl_bundle_adjust_lsqr::set_residual_scale

(

double

scale

)

[inline]

set the residual scale for the robust estimation.

Definition at line 118 of file vpgl_bundle_adjust_lsqr.h.

---

Member Data Documentation

vcl_vector<vnl_matrix<double> > vpgl_bundle_adjust_lsqr::factored_inv_covars_ [protected]

The Cholesky factored inverse covariances for each image point.

Definition at line 216 of file vpgl_bundle_adjust_lsqr.h.

vcl_vector<vgl_point_2d<double> > vpgl_bundle_adjust_lsqr::image_points_ [protected]

The corresponding points in the image.

Definition at line 214 of file vpgl_bundle_adjust_lsqr.h.

int vpgl_bundle_adjust_lsqr::iteration_count_ [protected]

Definition at line 222 of file vpgl_bundle_adjust_lsqr.h.

double vpgl_bundle_adjust_lsqr::scale2_ [protected]

The square of the scale of the robust estimator.

Definition at line 220 of file vpgl_bundle_adjust_lsqr.h.

bool vpgl_bundle_adjust_lsqr::use_covars_ [protected]

Flag to enable covariance weighted errors.

Definition at line 218 of file vpgl_bundle_adjust_lsqr.h.

---

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

• core/vpgl/algo/vpgl_bundle_adjust_lsqr.h
• core/vpgl/algo/vpgl_bundle_adjust_lsqr.cxx