SpinImageNormal Class Reference

A SpinImageNormal descriptor computes a spin image spinning around the locally extracted normal vector. More...

#include <spin_image_normal.h>

Inheritance diagram for SpinImageNormal:

List of all members.

Public Member Functions
virtual void	clearShared ()
	Clears any already-computed spectral information.
std::string	getName () const
	Returns a string that is unique for the current param settings.
	SpinImageNormal ()
	Instantiates the spin image descriptor to use the given specifications.
	SpinImageNormal (const double row_res, const double col_res, const unsigned int nbr_rows, const unsigned int nbr_cols, const bool use_interest_regions_only, SpectralAnalysis &spectral_information)
Protected Member Functions
virtual int	precompute (const sensor_msgs::PointCloud &data, cloud_kdtree::KdTree &data_kdtree, const std::vector< const geometry_msgs::Point32 * > &interest_pts)
	Computes/retrieves the normal for each interest point.
virtual int	precompute (const sensor_msgs::PointCloud &data, cloud_kdtree::KdTree &data_kdtree, const std::vector< const std::vector< int > * > &interest_region_indices)
	Computes/retrieves the normal for each interest region.
Private Attributes
SpectralAnalysis *	spectral_information_

---

Detailed Description

A SpinImageNormal descriptor computes a spin image spinning around the locally extracted normal vector.

The normal vector is the "beta" axis as described in Johnson & Hebert 1999.

Example spin image definition with 3 rows and 4 cols:
beta
^
|_ _ _ _
|_|_|_|_|
x_|_|_|_|
|_|_|_|_|
-----------> alpha
(x = center point of spin image, beta = [normal vector])

The center point of the spin image is the given interest point or the centroid of given regions of interest points

Definition at line 78 of file spin_image_normal.h.

---

Constructor & Destructor Documentation

SpinImageNormal::SpinImageNormal

(

)

[inline]

Instantiates the spin image descriptor to use the given specifications.

Warning:

The number of rows (nbr_rows) must be odd

Parameters:

row_res	The cell resolution along the beta axis
col_res	The cell resolution along the alpha axis
nbr_rows	The number of cells along the beta axis
nbr_cols	The number of cells along the alpha axis
use_interest_regions_only	When computing for interest regions, true indicates to use only the points within the interest region to compute the spin image
spectral_information	The class to retrieve the normal vectors from for each interest point/region

Definition at line 98 of file spin_image_normal.h.

SpinImageNormal::SpinImageNormal	(	const double	row_res,
		const double	col_res,
		const unsigned int	nbr_rows,
		const unsigned int	nbr_cols,
		const bool	use_interest_regions_only,
		SpectralAnalysis &	spectral_information
	)

Definition at line 42 of file spin_image_normal.cpp.

---

Member Function Documentation

void SpinImageNormal::clearShared

(

)

[virtual]

Clears any already-computed spectral information.

Implements Descriptor3D.

Definition at line 81 of file spin_image_normal.cpp.

string SpinImageNormal::getName

(

)

const [virtual]

Returns a string that is unique for the current param settings.

Returns:

the name of this feature.

Implements Descriptor3D.

Definition at line 72 of file spin_image_normal.cpp.

int SpinImageNormal::precompute	(	const sensor_msgs::PointCloud &	data,
		cloud_kdtree::KdTree &	data_kdtree,
		const std::vector< const geometry_msgs::Point32 * > &	interest_pts
	)		[protected, virtual]

Computes/retrieves the normal for each interest point.

Parameters:

data	The point cloud to process from Descriptor3D::compute()
data_kdtree	The efficient neighborhood data structure
interest_pts	The list of interest points to be processed

Returns:

0 on success, otherwise negative value on error

Implements Descriptor3D.

Definition at line 89 of file spin_image_normal.cpp.

int SpinImageNormal::precompute	(	const sensor_msgs::PointCloud &	data,
		cloud_kdtree::KdTree &	data_kdtree,
		const std::vector< const std::vector< int > * > &	interest_region_indices
	)		[protected, virtual]

Computes/retrieves the normal for each interest region.

Parameters:

data	The point cloud to process from Descriptor3D::compute()
data_kdtree	The efficient neighborhood data structure
interest_pts	The list of interest points to be processed

Returns:

0 on success, otherwise negative value on error

Implements Descriptor3D.

Definition at line 133 of file spin_image_normal.cpp.

---

Member Data Documentation

SpectralAnalysis* SpinImageNormal::spectral_information_ [private]

Definition at line 147 of file spin_image_normal.h.

---

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

• spin_image_normal.h
• spin_image_normal.cpp