#include <cloud.h>

Inheritance diagram for cloud3:

Public Member Functions
	cloud3 (const options &opts, const calibration &_calib)
void	init ()
virtual bool	read_background (const std::string &depth_background)
void	render_background_image (IplImage *retval, double scale=1.0) const
void	show () const
vector3	operator[] (const size_t i) const
size_t	size () const
const project2d &	get_project2d () const
const knn_cloud &	get_knn_cloud () const
const knn_cloud &	get_knn_bg_cloud () const
Protected Attributes
point_vector	points
point_vector	colours
const int	downsample
const int	first_frame
bool	use_background_subtraction
std::string	depth_background
project2d	proj2d
knn_cloud	nearest_cloud
knn_cloud	bg_nearest_cloud
IplImage *	background

Detailed Description

cloud3 is an abstract class used for representing observations consisting of coloured 3D points, such as those generated by stereo cameras or depth sensors.

Constructor & Destructor Documentation

cloud3::cloud3	(	const options &	opts,
		const calibration &	_calib
	)			`[inline]`

Constructor for the cloud3 class. Call this from derived class constructors. Also, do not forgot to call the 'init' function as part of the derived constructor.

Member Function Documentation

const knn_cloud& cloud3::get_knn_bg_cloud ( ) const [inline]

Return a knn_cloud object corresponding to the 3D points in the background.

Returns:: An object of class knn_cloud.

const knn_cloud& cloud3::get_knn_cloud ( ) const [inline]

Return a knn_cloud object corresponding to the 3D points in the point_cloud.

Returns:: An object of class knn_cloud.

const project2d& cloud3::get_project2d ( ) const [inline]

Return a project2d object where the binary segment corresponds to a projection of the 3D points onto the image plane. This allows for easy 2D projection of 2D points onto the binary segment.

Returns:: An object of class project2d. The used binary segment is a projection of the 3D points onto the original image plane.

void cloud3::init ( ) [inline]

Initialise the cloud3 object. You need to call this in the constructor of derived classes.

vector3 cloud3::operator[] ( const size_t i ) const [inline]

Return one 3D point. This allows you to access the data currently held in the object. As an example,

     for (size_t i = 0; i < my_point_cloud.size (); i++)
       do_something_with_point (my_point_cloud [i]);

Parameters:

[in] i The index of the point you want to read.

Returns:: The 3D point at the given index.

virtual bool cloud3::read_background ( const std::string & depth_background ) [inline, virtual]

Attempt to read information about the scene background. You need to implement this function in derived classes.

Returns:: true if the background could be read succesfully; false otherwise.

Reimplemented in kinect_cloud, point_cloud, and triclops_cloud.

void cloud3::render_background_image	(	IplImage *	retval,
		double	scale = `1.0`
	)			const `[inline]`

Render background subtracted image from the 3D point. Each 3D point currently stored in the object is projected into retval (using the calibration object given to the object constructor). Projected pixels are rendered with a value of 255, while other pixels are assigned a 0 value. A morphologic closing operation is performed on the image afterwards.

Parameters:

`[out]`	retval	An 8-bit unsigned gray scale image (IPL_DEPTH_8U with 1 channel) in which the result is rendered.
`[in]`	scale	An optional scaling parameter. This is passed to the world_to_pixel function in the calibration object.