Visual Servoing Platform  version 3.1.0
Public Types | Public Member Functions | Static Public Member Functions | Public Attributes | List of all members
vpCalibration Class Reference
vision: Computer vision module » Camera calibration

#include <visp3/vision/vpCalibration.h>

Public Types

enum  vpCalibrationMethodType {
  CALIB_LAGRANGE, CALIB_VIRTUAL_VS, CALIB_VIRTUAL_VS_DIST, CALIB_LAGRANGE_VIRTUAL_VS,
  CALIB_LAGRANGE_VIRTUAL_VS_DIST
}
 

Public Member Functions

 vpCalibration ()
 
 vpCalibration (const vpCalibration &c)
 
virtual ~vpCalibration ()
 
int addPoint (double X, double Y, double Z, vpImagePoint &ip)
 
vpCalibration & operator= (const vpCalibration &twinCalibration)
 
int clearPoint ()
 
void computeStdDeviation (double &deviation, double &deviation_dist)
 
int computeCalibration (vpCalibrationMethodType method, vpHomogeneousMatrix &cMo_est, vpCameraParameters &cam_est, bool verbose=false)
 
double computeStdDeviation (const vpHomogeneousMatrix &cMo_est, const vpCameraParameters &camera)
 
double computeStdDeviation_dist (const vpHomogeneousMatrix &cMo, const vpCameraParameters &cam)
 
int displayData (vpImage< unsigned char > &I, vpColor color=vpColor::red, unsigned int thickness=1, int subsampling_factor=1)
 
int displayGrid (vpImage< unsigned char > &I, vpColor color=vpColor::yellow, unsigned int thickness=1, int subsampling_factor=1)
 
double getResidual (void) const
 
double getResidual_dist (void) const
 
unsigned int get_npt () const
 
int init ()
 
int readData (const char *filename)
 
int writeData (const char *filename)
 

Static Public Member Functions

static void calibrationTsai (const std::vector< vpHomogeneousMatrix > &cMo, const std::vector< vpHomogeneousMatrix > &rMe, vpHomogeneousMatrix &eMc)
 
static int computeCalibrationMulti (vpCalibrationMethodType method, std::vector< vpCalibration > &table_cal, vpCameraParameters &cam, double &globalReprojectionError, bool verbose=false)
 
static int computeCalibrationTsai (const std::vector< vpCalibration > &table_cal, vpHomogeneousMatrix &eMc, vpHomogeneousMatrix &eMc_dist)
 
static double getLambda ()
 
static int readGrid (const char *filename, unsigned int &n, std::list< double > &oX, std::list< double > &oY, std::list< double > &oZ, bool verbose=false)
 
static void setLambda (const double &lambda)
 

Public Attributes

vpHomogeneousMatrix cMo
 
vpHomogeneousMatrix cMo_dist
 
vpCameraParameters cam
 
vpCameraParameters cam_dist
 
vpHomogeneousMatrix rMe
 
vpHomogeneousMatrix eMc
 
vpHomogeneousMatrix eMc_dist
 

Detailed Description

Tools for perspective camera calibration.

Definition at line 71 of file vpCalibration.h.

Member Enumeration Documentation

◆ vpCalibrationMethodType

enum vpCalibration::vpCalibrationMethodType

Minimization algorithm use to estimate the camera parameters.

Enumerator
CALIB_LAGRANGE 

Lagrange approach without estimation of the distortion.

CALIB_VIRTUAL_VS 

Virtual visual servoing approach without estimation of the distortion (results are similar to Lowe approach).

CALIB_VIRTUAL_VS_DIST 

Virtual visual servoing approach with estimation of the distortion.

CALIB_LAGRANGE_VIRTUAL_VS 

Lagrange approach first, than virtual visual servoing approach, without estimation of the distortion.

CALIB_LAGRANGE_VIRTUAL_VS_DIST 

Lagrange approach first, than virtual visual servoing approach, with estimation of the distortion.

Definition at line 77 of file vpCalibration.h.

Constructor & Destructor Documentation

◆ vpCalibration() [1/2]

vpCalibration::vpCalibration ( )

Default constructor.

Definition at line 74 of file vpCalibration.cpp.

References init().

◆ vpCalibration() [2/2]

vpCalibration::vpCalibration ( const vpCalibration &  c)

Copy constructor.

Definition at line 83 of file vpCalibration.cpp.

◆ ~vpCalibration()

vpCalibration::~vpCalibration ( )
virtual

Destructor : delete the array of point (freed the memory)

Definition at line 93 of file vpCalibration.cpp.

References clearPoint().

Member Function Documentation

◆ addPoint()

int vpCalibration::addPoint ( double  X,
double  Y,
double  Z,
vpImagePoint &  ip 
)

Add a new point in the array of points.

Parameters
X,Y,Z: 3D coordinates of a point in the object frame
ip: 2D Coordinates of the point in the camera frame.

Definition at line 144 of file vpCalibration.cpp.

References vpPose::addPoint(), vpPose::clearPoint(), vpPose::computePose(), vpPose::computeResidual(), vpPixelMeterConversion::convertPoint(), vpPose::DEMENTHON, vpPose::LAGRANGE, vpPoint::set_x(), vpPoint::set_y(), and vpPose::VIRTUAL_VS.

Referenced by readData().

◆ calibrationTsai()

void vpCalibration::calibrationTsai ( const std::vector< vpHomogeneousMatrix > &  cMo,
const std::vector< vpHomogeneousMatrix > &  rMe,
vpHomogeneousMatrix &  eMc 
)
static

calibration method of effector-camera from R. Tsai and R. Lorenz [41].

Compute extrinsic camera parameters : the constant transformation from the effector to the camera coordinates (eMc).

Parameters
cMo: vector of homogeneous matrices representing the transformation between the camera and the scene (input)
rMe: vector of homogeneous matrices representing the transformation between the effector (where the camera is fixed) and the reference coordinates (base of the manipulator) (input). Must be the same size as cMo.
eMc: homogeneous matrix representing the transformation between the effector and the camera (output)
Examples:
calibrateTsai.cpp.

Definition at line 1069 of file vpCalibrationTools.cpp.

References vpMatrix::AtA(), cam, cam_dist, cMo, cMo_dist, computeStdDeviation(), computeStdDeviation_dist(), vpCalibrationException::convergencyError, vpException::dimensionError, vpExponentialMap::direct(), vpMath::equal(), vpRotationMatrix::eye(), vpCameraParameters::get_kdu(), vpCameraParameters::get_kud(), vpCameraParameters::get_px(), vpCameraParameters::get_py(), vpImagePoint::get_u(), vpCameraParameters::get_u0(), vpImagePoint::get_v(), vpCameraParameters::get_v0(), vpCameraParameters::initPersProjWithDistortion(), vpCameraParameters::initPersProjWithoutDistortion(), vpHomogeneousMatrix::insert(), vpHomogeneousMatrix::inverse(), vpCalibrationException::notInitializedError, vpMatrix::pseudoInverse(), vpMath::sinc(), vpArray2D< Type >::size(), vpColVector::skew(), vpMath::sqr(), vpMatrix::stack(), vpColVector::stack(), vpColVector::sumSquare(), vpRotationMatrix::t(), vpMatrix::t(), and vpERROR_TRACE.

Referenced by computeCalibrationTsai().

◆ clearPoint()

int vpCalibration::clearPoint ( )

Suppress all the point in the array of point.

Delete the array of points.

Definition at line 127 of file vpCalibration.cpp.

Referenced by readData(), and ~vpCalibration().

◆ computeCalibration()

int vpCalibration::computeCalibration ( vpCalibrationMethodType  method,
vpHomogeneousMatrix &  cMo_est,
vpCameraParameters &  cam_est,
bool  verbose = false 
)

Compute the calibration according to the desired method using one pose.

Parameters
method: Method that will be used to estimate the parameters.
cMo_est: estimated homogeneous matrix that defines the pose.
cam_est: estimated intrinsic camera parameters.
verbose: set at true if information about the residual at each loop of the algorithm is hoped.
Returns
0 if the calibration computation succeed.

Definition at line 356 of file vpCalibration.cpp.

References CALIB_LAGRANGE, CALIB_LAGRANGE_VIRTUAL_VS, CALIB_LAGRANGE_VIRTUAL_VS_DIST, CALIB_VIRTUAL_VS, CALIB_VIRTUAL_VS_DIST, cam, cam_dist, cMo, cMo_dist, and vpCameraParameters::printParameters().

◆ computeCalibrationMulti()

int vpCalibration::computeCalibrationMulti ( vpCalibrationMethodType  method,
std::vector< vpCalibration > &  table_cal,
vpCameraParameters &  cam_est,
double &  globalReprojectionError,
bool  verbose = false 
)
static

Compute the multi-images calibration according to the desired method using many poses.

Parameters
method: Method used to estimate the camera parameters.
table_cal: Vector of vpCalibration.
cam_est: Estimated intrinsic camera parameters.
globalReprojectionError: Global reprojection error or global residual.
verbose: Set at true if information about the residual at each loop of the algorithm is hoped.
Returns
0 if the computation was managed succeed.

Definition at line 445 of file vpCalibration.cpp.

References CALIB_LAGRANGE, CALIB_LAGRANGE_VIRTUAL_VS, CALIB_LAGRANGE_VIRTUAL_VS_DIST, CALIB_VIRTUAL_VS, CALIB_VIRTUAL_VS_DIST, cMo, get_npt(), and vpCameraParameters::printParameters().

◆ computeCalibrationTsai()

int vpCalibration::computeCalibrationTsai ( const std::vector< vpCalibration > &  table_cal,
vpHomogeneousMatrix &  eMc,
vpHomogeneousMatrix &  eMc_dist 
)
static

Compute the multi-image calibration of effector-camera from R. Tsai and R. Lenz [41].

Compute extrinsic camera parameters : the constant transformation from the end-effector to the camera frame ${^e}{\bf M}_c$ considering the camera model with or without distortion.

Parameters
[in]table_cal: Vector of vpCalibration that contains for each index a couple of ${^r}{\bf M}_e$ (world to end-effector) and ${^c}{\bf M}_o$ (camera to object) transformations.
[out]eMc: Estimated pose of the camera in relation to the end-effector considering the camera model without distortion.
[out]eMc_dist: Estimated pose of the camera in relation to the end-effector considering the model with distortion.
Returns
0 if the computation managed, -1 if less than three poses are provides as input.

Definition at line 543 of file vpCalibration.cpp.

References calibrationTsai(), and vpERROR_TRACE.

◆ computeStdDeviation() [1/2]

void vpCalibration::computeStdDeviation ( double &  deviation,
double &  deviation_dist 
)

Compute and return the standard deviation expressed in pixel for pose matrix and camera intrinsic parameters.

Parameters
deviation: the standard deviation computed for the model without distortion.
deviation_dist: the standard deviation computed for the model with distortion.

Definition at line 339 of file vpCalibration.cpp.

References cam, cam_dist, cMo, cMo_dist, and computeStdDeviation_dist().

Referenced by calibrationTsai().

◆ computeStdDeviation() [2/2]

double vpCalibration::computeStdDeviation ( const vpHomogeneousMatrix &  cMo_est,
const vpCameraParameters &  camera 
)

Compute and return the standard deviation expressed in pixel for pose matrix and camera intrinsic parameters for model without distortion.

Parameters
cMo_est: the matrix that defines the pose to be tested.
camera: camera intrinsic parameters to be tested.
Returns
the standard deviation by point of the error in pixel .

Definition at line 221 of file vpCalibration.cpp.

References vpCameraParameters::get_px(), vpCameraParameters::get_py(), vpImagePoint::get_u(), vpCameraParameters::get_u0(), vpImagePoint::get_v(), vpCameraParameters::get_v0(), and vpMath::sqr().

◆ computeStdDeviation_dist()

double vpCalibration::computeStdDeviation_dist ( const vpHomogeneousMatrix &  cMo_est,
const vpCameraParameters &  camera 
)

Compute and return the standard deviation expressed in pixel for pose matrix and camera intrinsic parameters with pixel to meter model.

Parameters
cMo_est: the matrix that defines the pose to be tested.
camera: camera intrinsic parameters to be tested.
Returns
the standard deviation by point of the error in pixel .

Definition at line 272 of file vpCalibration.cpp.

References vpCameraParameters::get_kdu(), vpCameraParameters::get_kud(), vpCameraParameters::get_px(), vpCameraParameters::get_py(), vpImagePoint::get_u(), vpCameraParameters::get_u0(), vpImagePoint::get_v(), vpCameraParameters::get_v0(), and vpMath::sqr().

Referenced by calibrationTsai(), and computeStdDeviation().

◆ displayData()

int vpCalibration::displayData ( vpImage< unsigned char > &  I,
vpColor  color = vpColor::red,
unsigned int  thickness = 1,
int  subsampling_factor = 1 
)

Display the data of the calibration (center of the tracked dots)

Parameters
I: Image where to display data.
color: Color of the data.
thickness: Thickness of the displayed data.
subsampling_factor: Subsampling factor. Default value is 1. Admissible values are multiple of 2. Divide by this parameter the coordinates of the data points resulting from image processing.

Definition at line 719 of file vpCalibration.cpp.

References vpDisplay::displayCross(), vpImagePoint::get_u(), vpImagePoint::get_v(), vpImagePoint::set_u(), and vpImagePoint::set_v().

◆ displayGrid()

int vpCalibration::displayGrid ( vpImage< unsigned char > &  I,
vpColor  color = vpColor::yellow,
unsigned int  thickness = 1,
int  subsampling_factor = 1 
)

Display estimated centers of dots using intrinsic camera parameters with model with distortion and the computed pose.

Parameters
I: Image where to display grid data.
color: Color of the data.
thickness: Thickness of the displayed data.
subsampling_factor: Subsampling factor. Default value is 1. Admissible values are multiple of 2. Divide by this parameter the values of the camera parameters.

Definition at line 743 of file vpCalibration.cpp.

References cam_dist, cMo_dist, vpDisplay::displayCross(), vpCameraParameters::get_kud(), vpCameraParameters::get_px(), vpCameraParameters::get_py(), vpCameraParameters::get_u0(), vpCameraParameters::get_v0(), vpImagePoint::set_u(), and vpMath::sqr().

◆ get_npt()

unsigned int vpCalibration::get_npt ( ) const
inline

get the number of points

Definition at line 151 of file vpCalibration.h.

Referenced by computeCalibrationMulti().

◆ getLambda()

static double vpCalibration::getLambda ( )
inlinestatic

Set the gain for the virtual visual servoing algorithm.

Definition at line 144 of file vpCalibration.h.

◆ getResidual()

double vpCalibration::getResidual ( void  ) const
inline

get the residual in pixels

Definition at line 147 of file vpCalibration.h.

◆ getResidual_dist()

double vpCalibration::getResidual_dist ( void  ) const
inline

get the residual for perspective projection with distortion (in pixels)

Definition at line 149 of file vpCalibration.h.

◆ init()

int vpCalibration::init ( void  )

Basic initialisation (called by the constructors)

Definition at line 57 of file vpCalibration.cpp.

Referenced by vpCalibration().

◆ operator=()

vpCalibration & vpCalibration::operator= ( const vpCalibration &  twinCalibration)

= operator.

Parameters
twinCalibration: object to be copied

Definition at line 100 of file vpCalibration.cpp.

References cam, cam_dist, cMo, cMo_dist, eMc, eMc_dist, and rMe.

◆ readData()

int vpCalibration::readData ( const char *  filename)

Read data from disk : data are organized as follow oX oY oZ u v

Parameters
filename: name of the file

Definition at line 620 of file vpCalibration.cpp.

References addPoint(), vpException::badValue, clearPoint(), vpImagePoint::set_u(), and vpImagePoint::set_v().

◆ readGrid()

int vpCalibration::readGrid ( const char *  filename,
unsigned int &  n,
std::list< double > &  oX,
std::list< double > &  oY,
std::list< double > &  oZ,
bool  verbose = false 
)
static

Read calibration grid coordinates from disk : data are organized as follow oX oY oZ

Parameters
filename: name of the file
n: number of points in the calibration grid
oX: List of oX coordinates
oY: List of oY coordinates
oZ: List of oZ coordinates
verbose: Additionnal printings if true (number of points on the calibration grid and their respective coordinates in the object frame).
Returns
0 if success, -1 if an error occurs.

Definition at line 666 of file vpCalibration.cpp.

References vpException::badValue.

◆ setLambda()

static void vpCalibration::setLambda ( const double &  lambda)
inlinestatic

set the gain for the virtual visual servoing algorithm

Definition at line 160 of file vpCalibration.h.

◆ writeData()

int vpCalibration::writeData ( const char *  filename)

Write data into a file.

data are organized as follow oX oY oZ u v

Parameters
filename: name of the file

Definition at line 578 of file vpCalibration.cpp.

References vpImagePoint::get_u(), and vpImagePoint::get_v().

Member Data Documentation

◆ cam

vpCameraParameters vpCalibration::cam

projection model without distortion

camera intrinsic parameters for perspective

Definition at line 98 of file vpCalibration.h.

Referenced by calibrationTsai(), computeCalibration(), computeStdDeviation(), and operator=().

◆ cam_dist

vpCameraParameters vpCalibration::cam_dist

projection model with distortion

camera intrinsic parameters for perspective

Definition at line 100 of file vpCalibration.h.

Referenced by calibrationTsai(), computeCalibration(), computeStdDeviation(), displayGrid(), and operator=().

◆ cMo

vpHomogeneousMatrix vpCalibration::cMo

(as a 3x4 matrix [R T])

the pose computed for the model without distortion

Definition at line 93 of file vpCalibration.h.

Referenced by calibrationTsai(), computeCalibration(), computeCalibrationMulti(), computeStdDeviation(), and operator=().

◆ cMo_dist

vpHomogeneousMatrix vpCalibration::cMo_dist

the pose computed for perspective projection with distortion model (as a 3x4 matrix [R T])

Definition at line 95 of file vpCalibration.h.

Referenced by calibrationTsai(), computeCalibration(), computeStdDeviation(), displayGrid(), and operator=().

◆ eMc

vpHomogeneousMatrix vpCalibration::eMc

position of the camera in relation to the effector

Definition at line 105 of file vpCalibration.h.

Referenced by operator=().

◆ eMc_dist

vpHomogeneousMatrix vpCalibration::eMc_dist

Definition at line 106 of file vpCalibration.h.

Referenced by operator=().

◆ rMe

vpHomogeneousMatrix vpCalibration::rMe

reference coordinates (manipulator base coordinates)

position of the effector in relation to the

Definition at line 103 of file vpCalibration.h.

Referenced by operator=().