base::samples::DistanceImage Struct Reference

#include <DistanceImage.hpp>

Public Types
typedef float	scalar

Public Member Functions
	DistanceImage ()
	DistanceImage (uint16_t width, uint16_t height)
void	clear ()
template<class Scalar_ >
bool	getScenePoint (size_t x, size_t y, Eigen::Matrix< Scalar_, 3, 1 > &point) const
template<class Scalar_ >
bool	getImagePoint (const Eigen::Matrix< Scalar_, 3, 1 > &point, size_t &x, size_t &y) const
base::samples::Pointcloud	getPointCloud () const
template<class Scalar_ >
Eigen::Matrix< Scalar_, 3, 3 >	getIntrinsic () const
void	setIntrinsic (double f_x, double f_y, double c_x, double c_y)
	set the intrinsic camera parameters. More...
void	setSize (uint16_t width, uint16_t height)
	set the size of the distance image More...

Public Attributes
base::Time	time
	original timestamp of the camera image More...
uint16_t	width
	width (x) value in pixels More...
uint16_t	height
	height (y) value in pixels More...
scalar	scale_x
	scale value to apply to the x axis More...
scalar	scale_y
	scale value to apply to the y axis More...
scalar	center_x
	center offset to apply to the x axis More...
scalar	center_y
	center offset to apply to the y axis More...
std::vector< scalar >	data
	distance values stored in row major order. NaN is used as the no value type. More...

Detailed Description

2D array structure representing a distance image for a pinhole camera model.

The grid pixels are scaled such that (x*scale_x)+center_x = p_x are the projective plane coordinates given a grid index x. This of course applies to y as well.

The data array is a row major flattened version of the image matrix, giving the distance value d of the image points. The relation is such that for a point on the projection plane, the 3D point z can be calculated as (p_x,p_y,1)*d = z.

NOTE: unfortunately the scale and center is directly inverse to the convention used in: http://docs.opencv.org/modules/calib3d/doc/camera_calibration_and_3d_reconstruction.html for the time being, scale_x and scale_y are kept public for legacy code. It is recommended to use the functions operating on it instead.

Member Typedef Documentation

scalar

typedef float base::samples::DistanceImage::scalar

Constructor & Destructor Documentation

DistanceImage() [1/2]

base::samples::DistanceImage::DistanceImage ( )

inline

DistanceImage() [2/2]

base::samples::DistanceImage::DistanceImage ( uint16_t  width, uint16_t  height  )

inline

Member Function Documentation

clear()

void base::samples::DistanceImage::clear

(

)

resets all values in the distance image to nan

getImagePoint()

template<class Scalar_ >

bool base::samples::DistanceImage::getImagePoint ( const Eigen::Matrix< Scalar_, 3, 1 > &  point, size_t &  x, size_t &  y  ) const

inline

transforms a 3d scene point onto an image coordinate

Parameters

point	- [in] 3d-scene point
x	- [out] x image coordinate
y	- [out] y image coordinate

getIntrinsic()

template<class Scalar_ >

Eigen::Matrix<Scalar_,3,3> base::samples::DistanceImage::getIntrinsic ( ) const

inline

The intrinsic matrix has the following form

| f_x 0 c_x | | 0 f_y c_y | | 0 0 1 |

where f_x and f_y are the focal length expressed in pixel and c_x and c_y are the center in pixel. This matrix can be used to convert screen coordinates (x, y, z) into homogenous image coordinates (u, v, w) in the following way:

(u,v,w) = F * (x, y, z)

Note, that f and c are the inverse of the center and scale parameters.

Returns

the intrinsic camera matrix

getPointCloud()

Pointcloud base::samples::DistanceImage::getPointCloud

(

)

const

getScenePoint()

template<class Scalar_ >

bool base::samples::DistanceImage::getScenePoint ( size_t  x, size_t  y, Eigen::Matrix< Scalar_, 3, 1 > &  point  ) const

inline

Transforms a x, y pixel coordinates into a 3d scene point, using the distance value of that point if available

Parameters

x	- [in] x coordinate
y	- [in] y coordinate
point	- [out] resulting 3d scene point

Returns

true if the scene point has a corresponding distance value

setIntrinsic()

void base::samples::DistanceImage::setIntrinsic	(	double	f_x,
		double	f_y,
		double	c_x,
		double	c_y
	)

set the intrinsic camera parameters.

see getIntrinsic() for a more detailed explanation.

Parameters

f_x	focal length in x
f_y	focal length in y
c_x	center point in x
c_y	center point in y

setSize()

void base::samples::DistanceImage::setSize	(	uint16_t	width,
		uint16_t	height
	)

set the size of the distance image

This will also rescale the internal data array to hold the appropriate number of pixels.

Parameters

width	width of the image
height	height of the image

Member Data Documentation

center_x

scalar base::samples::DistanceImage::center_x

center offset to apply to the x axis

center_y

scalar base::samples::DistanceImage::center_y

center offset to apply to the y axis

data

std::vector<scalar> base::samples::DistanceImage::data

distance values stored in row major order. NaN is used as the no value type.

height

uint16_t base::samples::DistanceImage::height

height (y) value in pixels

scale_x

scalar base::samples::DistanceImage::scale_x

scale value to apply to the x axis

scale_y

scalar base::samples::DistanceImage::scale_y

scale value to apply to the y axis

time

base::Time base::samples::DistanceImage::time

original timestamp of the camera image

width

uint16_t base::samples::DistanceImage::width

width (x) value in pixels

---

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

• /home/dfki.uni-bremen.de/roehr/workspace/2maz/dev/base/types/src/samples/DistanceImage.hpp
• /home/dfki.uni-bremen.de/roehr/workspace/2maz/dev/base/types/src/samples/DistanceImage.cpp