HVR::DepthCamera Class Reference

Depth camera is the base type for all nodes that provide depth camera data, from such devices as the Kinect or Primesense Sensor. More...

Inheritance diagram for HVR::DepthCamera:

Classes
struct	ExtremesMap
	A map of extremes. More...
struct	SFBackgroundImage
struct	SFPositionImage
struct	Shape
struct	ViewSize

Public Types
typedef H3D::H3DImageObject::SFImage	SFImage
typedef H3D::TypedSFNode < Candy::PaintTexture >	SFPaintTexture
typedef std::vector< H3D::Vec3f >	contour_t
typedef std::vector< contour_t >	contours_t

Public Member Functions
	DepthCamera (H3D::Inst< H3D::SFNode > _metadata=0, H3D::Inst< H3D::SFVec2f > _fieldOfView=0, H3D::Inst< ViewSize > _viewSize=0, H3D::Inst< H3D::SFBool > _updateBackground=0, H3D::Inst< H3D::SFFloat > _maximumBackgroundDistance=0, H3D::Inst< H3D::SFFloat > _backgroundDepthMargin=0, H3D::Inst< H3D::SFInt32 > _backgroundPixelMargin=0, H3D::Inst< SFImage > _depthData=0, H3D::Inst< SFPositionImage > _positionData=0, H3D::Inst< SFBackgroundImage > _depthDataBackground=0, H3D::Inst< SFPaintTexture > _debugTexture=0)
virtual std::string	defaultXMLContainerField ()
H3D::Vec3f	pixel2DtoLine (float x, float y)
	Convert a pixel coordinate on the Kinect image into a 3D line representing that pixels direction from the Kinects point of view. More...
H3D::Vec3f	pixel2Dto3D (float x, float y, float d)
	Convert a pixel coordinate on the Kinect image and depth value into a 3D position from the Kinects point of view. More...
H3D::Vec3f	pixel2Dto3D (H3D::Vec3f p)
	Convert a pixel coordinate on the Kinect image (x,y,depth) into a 3D position from the Kinects point of view. More...
H3D::Vec3f	point2DtoLine (float x, float y)
	Convert a point on the Kinect image into a 3D line representing that pixels direction from the Kinects point of view. More...
H3D::Vec3f	point2Dto3D (float x, float y, float d)
	Convert a point on the Kinect image into a 3D position from the Kinects point of view. More...
H3D::Vec3f	point2Dto3D (H3D::Vec3f p)
	Convert a point on the Kinect image into a 3D position from the Kinects point of view. More...
const std::vector< Shape > &	getShapes ()
H3D::Vec2f	getExtremes ()
	Returns a 2D vector containing the largest and smallest value of the depth in x and y, respectively. More...
const ExtremesMap &	getExtremesMap ()
	Returns a map of extremes. More...

Public Attributes
struct HVR_API	Shape
struct HVR_API	ExtremesMap
std::auto_ptr< H3D::SFVec2f >	fieldOfView
	This camera field of view is used when converting coordinates between 2D and 3D. More...
std::auto_ptr< ViewSize >	viewSize
	This field holds the width and height of the view at the distance of one length unit. More...
std::auto_ptr< H3D::SFBool >	updateBackground
	The background image (depthDataBackground) will be updated if this field is set to true. More...
std::auto_ptr< H3D::SFFloat >	maximumBackgroundDistance
	This is the largest distance that the background image may contain. More...
std::auto_ptr< H3D::SFFloat >	backgroundDepthMargin
	This is a margin removed from the depth when specifying the background, in meters. More...
std::auto_ptr< H3D::SFInt32 >	backgroundPixelMargin
	This is the margin around object when specifying the background, in pixels. More...
std::auto_ptr< SFImage >	depthData
	This is the depth data from the depth camera, in meters. More...
std::auto_ptr< SFPositionImage >	positionData
	This is the 3D pixel position data from the depth camera, in meters and camera coordinates. More...
std::auto_ptr< SFBackgroundImage >	depthDataBackground
	The depth camera holds a background image to more easily remove uninteresting data. More...
std::auto_ptr< SFPaintTexture >	debugTexture
	When specified, debug data are drawn to this texture. More...

Static Public Attributes
static H3D::H3DNodeDatabase	database

Protected Member Functions
void	update_debug_texture ()

Protected Attributes
struct HVR_API	ViewSize
struct HVR_API	SFBackgroundImage
struct HVR_API	SFPositionImage

Detailed Description

Depth camera is the base type for all nodes that provide depth camera data, from such devices as the Kinect or Primesense Sensor.

Member Function Documentation

H3D::Vec2f HVR::DepthCamera::getExtremes ( )

inline

Returns a 2D vector containing the largest and smallest value of the depth in x and y, respectively.

If there are no extremes, for example if everything is background, then the vector will be (-1,-1).

const ExtremesMap& HVR::DepthCamera::getExtremesMap ( )

inline

Returns a map of extremes.

Each cell of the map has a 2D vector containing the largest and smallest value of the depth for that cell in x and y, respectively. If there are no extremes of the cell, for example if the cell shows only background, then the vector will be (-1,-1).

H3D::Vec3f HVR::DepthCamera::pixel2Dto3D	(	float	x,
		float	y,
		float	d
	)

Convert a pixel coordinate on the Kinect image and depth value into a 3D position from the Kinects point of view.

This function will return origin if the actual 3D position could not be calculated.

H3D::Vec3f HVR::DepthCamera::pixel2Dto3D ( H3D::Vec3f  p )

inline

Convert a pixel coordinate on the Kinect image (x,y,depth) into a 3D position from the Kinects point of view.

The unit of depth is meters.

H3D::Vec3f HVR::DepthCamera::pixel2DtoLine	(	float	x,
		float	y
	)

Convert a pixel coordinate on the Kinect image into a 3D line representing that pixels direction from the Kinects point of view.

Returns a unit vector, or a zero vector if no depth image is available yet.

H3D::Vec3f HVR::DepthCamera::point2Dto3D	(	float	x,
		float	y,
		float	d
	)

Convert a point on the Kinect image into a 3D position from the Kinects point of view.

The x and y of the image point are in the range -1 to 1, while the depth is in meters.

H3D::Vec3f HVR::DepthCamera::point2Dto3D ( H3D::Vec3f  p )

inline

Convert a point on the Kinect image into a 3D position from the Kinects point of view.

The x and y of the image point are in the range -1 to 1, while z is in meters.

H3D::Vec3f HVR::DepthCamera::point2DtoLine	(	float	x,
		float	y
	)

Convert a point on the Kinect image into a 3D line representing that pixels direction from the Kinects point of view.

The image point is expressed in the range -1 to 1.

Member Data Documentation

std::auto_ptr< H3D::SFFloat > HVR::DepthCamera::backgroundDepthMargin

This is a margin removed from the depth when specifying the background, in meters.

This is a means to minimize the problem with moving background and poor depth accuracy.

std::auto_ptr< H3D::SFInt32 > HVR::DepthCamera::backgroundPixelMargin

This is the margin around object when specifying the background, in pixels.

This is a means to minimize the problem with moving background and poor depth accuracy.

std::auto_ptr< SFPaintTexture > HVR::DepthCamera::debugTexture

When specified, debug data are drawn to this texture.

std::auto_ptr< SFImage > HVR::DepthCamera::depthData

This is the depth data from the depth camera, in meters.

This can be set manually to emulate a depth camera but any extending node will set this field automatically. This must be an image with a single floating point channel. The image shows the camera's point of view and z being the distance from the camera, i e expressing a left-handed system. A negative value represents unavailable data.

std::auto_ptr< SFBackgroundImage > HVR::DepthCamera::depthDataBackground

The depth camera holds a background image to more easily remove uninteresting data.

This is an image with a single floating point channel.

std::auto_ptr< H3D::SFVec2f > HVR::DepthCamera::fieldOfView

This camera field of view is used when converting coordinates between 2D and 3D.

The camera sets this as good as it can but it may be manually specified.

std::auto_ptr< H3D::SFFloat > HVR::DepthCamera::maximumBackgroundDistance

This is the largest distance that the background image may contain.

std::auto_ptr< SFPositionImage > HVR::DepthCamera::positionData

This is the 3D pixel position data from the depth camera, in meters and camera coordinates.

std::auto_ptr< H3D::SFBool > HVR::DepthCamera::updateBackground

The background image (depthDataBackground) will be updated if this field is set to true.

The field will be set to false again when the background image has been updated.

std::auto_ptr< ViewSize > HVR::DepthCamera::viewSize

This field holds the width and height of the view at the distance of one length unit.

This is calculated by 2 x atan(fov/2).