A BoundingBoxSpectral descriptor to compute the dimensions of the bounding box in principle component space that encloses a neighborhood of points within a radius of the interest point/region. More...

#include <bounding_box_spectral.h>

Inheritance diagram for BoundingBoxSpectral:

[legend]

List of all members.

Public Member Functions
	BoundingBoxSpectral (double bbox_radius, SpectralAnalysis &spectral_information)
	Instantiates the bounding box feature with the specified radius to define the neighborhood and spectral information.
virtual void	clearShared ()
	Clears any already-computed spectral information.
std::string	getName () const
	Returns a name that is unique for any given setting of the parameters.
Protected Member Functions
virtual void	computeNeighborhoodFeature (const sensor_msgs::PointCloud &data, const std::vector< int > &neighbor_indices, const unsigned int interest_sample_idx, std::vector< float > &result) const
	Projects the given neighborhood in principle component space and then computes its bounding box.
virtual int	precompute (const sensor_msgs::PointCloud &data, cloud_kdtree::KdTree &data_kdtree, const std::vector< const geometry_msgs::Point32 * > &interest_pts)
	Computes the spectral information (eigenvectors), necessary to project the points into principle component space.
virtual int	precompute (const sensor_msgs::PointCloud &data, cloud_kdtree::KdTree &data_kdtree, const std::vector< const std::vector< int > * > &interest_region_indices)
	Computes the spectral information (eigenvectors), necessary to project the points into principle component space.
Private Attributes
const std::vector< const Eigen3::Vector3d * > *	eig_vecs_max_
	The biggest eigenvector for each interest point/region.
const std::vector< const Eigen3::Vector3d * > *	eig_vecs_mid_
	The middle eigenvector for each interest point/region.
const std::vector< const Eigen3::Vector3d * > *	eig_vecs_min_
	The smallest eigenvector for each interest point/region.
SpectralAnalysis *	spectral_information_
	The container the holds spectral information for the point cloud.

Detailed Description

A BoundingBoxSpectral descriptor to compute the dimensions of the bounding box in principle component space that encloses a neighborhood of points within a radius of the interest point/region.

The compute features are in order: [a,b,c] where a is the length along the principle eigenvector, b is the length along the middle eigenvector, and c is the length along the smallest eigenvector.

Definition at line 70 of file bounding_box_spectral.h.

Constructor & Destructor Documentation

BoundingBoxSpectral::BoundingBoxSpectral	(	double	bbox_radius,
		SpectralAnalysis &	spectral_information
	)

Instantiates the bounding box feature with the specified radius to define the neighborhood and spectral information.

When computing the feature for an interest region of points, the bounding box can either be the box that encloses the given region of points (indicated by -negative value), or from the neighboring points within the specified radius from the region's centroid (indicated by positive value).

Parameters:

bbox_radius	The radius from the interest point/region to define the neighborhood that defines the bounding box
spectral_information	Class to retrieve spectral information for the point cloud during Descriptor3D::compute()

Definition at line 42 of file bounding_box_spectral.cpp.

Member Function Documentation

void BoundingBoxSpectral::clearShared ( ) [virtual]

Clears any already-computed spectral information.

Implements Descriptor3D.

Definition at line 59 of file bounding_box_spectral.cpp.

void BoundingBoxSpectral::computeNeighborhoodFeature	(	const sensor_msgs::PointCloud &	data,
		const std::vector< int > &	neighbor_indices,
		const unsigned int	interest_sample_idx,
		std::vector< float > &	result
	)		const `[protected, virtual]`

Projects the given neighborhood in principle component space and then computes its bounding box.

Parameters:

data	The point cloud to process from Descriptor3D::compute()
neighbor_indices	The list of indices in data that constitute the neighborhood
interest_sample_idx	The index of the interest point/region that is being processed from Descriptor3D::compute()
result	The vector to hold the computed bounding box dimensions

Implements NeighborhoodFeature.

Definition at line 148 of file bounding_box_spectral.cpp.

std::string BoundingBoxSpectral::getName ( void ) const [virtual]

Returns a name that is unique for any given setting of the parameters.

Implements Descriptor3D.

Definition at line 67 of file bounding_box_spectral.cpp.

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

Computes the spectral information (eigenvectors), necessary to project the points into principle component space.

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 78 of file bounding_box_spectral.cpp.

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

Computes the spectral information (eigenvectors), necessary to project the points into principle component space.

Parameters:

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