Point Cloud Library (PCL)  1.14.0-dev
List of all members | Public Types | Public Member Functions | Protected Attributes
pcl::PointRepresentation< PointT > Class Template Referenceabstract

PointRepresentation provides a set of methods for converting a point structs/object into an n-dimensional vector. More...

#include <pcl/point_representation.h>

+ Inheritance diagram for pcl::PointRepresentation< PointT >:

Public Types

using Ptr = shared_ptr< PointRepresentation< PointT > >
 
using ConstPtr = shared_ptr< const PointRepresentation< PointT > >
 

Public Member Functions

virtual ~PointRepresentation ()=default
 Empty destructor. More...
 
virtual void copyToFloatArray (const PointT &p, float *out) const =0
 Copy point data from input point to a float array. More...
 
bool isTrivial () const
 Returns whether this point representation is trivial. More...
 
virtual bool isValid (const PointT &p) const
 Verify that the input point is valid. More...
 
template<typename OutputType >
void vectorize (const PointT &p, OutputType &out) const
 Convert input point into a vector representation, rescaling by alpha. More...
 
void vectorize (const PointT &p, float *out) const
 
void vectorize (const PointT &p, std::vector< float > &out) const
 
void setRescaleValues (const float *rescale_array)
 Set the rescale values to use when vectorizing points. More...
 
int getNumberOfDimensions () const
 Return the number of dimensions in the point's vector representation. More...
 

Protected Attributes

int nr_dimensions_ = 0
 The number of dimensions in this point's vector (i.e. More...
 
std::vector< float > alpha_
 A vector containing the rescale factor to apply to each dimension. More...
 
bool trivial_ = false
 Indicates whether this point representation is trivial. More...
 

Detailed Description

template<typename PointT>
class pcl::PointRepresentation< PointT >

PointRepresentation provides a set of methods for converting a point structs/object into an n-dimensional vector.

Note
This is an abstract class. Subclasses must set nr_dimensions_ to the appropriate value in the constructor and provide an implementation of the pure virtual copyToFloatArray method.
Author
Michael Dixon

Definition at line 59 of file point_representation.h.

Member Typedef Documentation

◆ ConstPtr

template<typename PointT >
using pcl::PointRepresentation< PointT >::ConstPtr = shared_ptr<const PointRepresentation<PointT> >

Definition at line 79 of file point_representation.h.

◆ Ptr

template<typename PointT >
using pcl::PointRepresentation< PointT >::Ptr = shared_ptr<PointRepresentation<PointT> >

Definition at line 78 of file point_representation.h.

Constructor & Destructor Documentation

◆ ~PointRepresentation()

template<typename PointT >
virtual pcl::PointRepresentation< PointT >::~PointRepresentation ( )
virtualdefault

Empty destructor.

Member Function Documentation

◆ copyToFloatArray()

template<typename PointT >
virtual void pcl::PointRepresentation< PointT >::copyToFloatArray ( const PointT p,
float *  out 
) const
pure virtual

Copy point data from input point to a float array.

This method must be overridden in all subclasses.

Parameters
[in]pThe input point
[out]outA pointer to a float array.

Implemented in pcl::DefaultFeatureRepresentation< VFHSignature308 >, pcl::DefaultPointRepresentation< UniqueShapeContext1960 >, pcl::DefaultPointRepresentation< SHOT352 >, pcl::DefaultPointRepresentation< SHOT1344 >, pcl::DefaultPointRepresentation< ShapeContext1980 >, pcl::DefaultPointRepresentation< PPFSignature >, pcl::DefaultFeatureRepresentation< PPFSignature >, pcl::GeneralizedIterativeClosestPoint6D::MyPointRepresentation, pcl::DefaultPointRepresentation< PointXYZI >, pcl::DefaultPointRepresentation< PointXYZ >, pcl::registration::TransformationValidationEuclidean< PointSource, PointTarget, Scalar >::MyPointRepresentation, pcl::DefaultPointRepresentation< PointNormal >, pcl::CustomPointRepresentation< PointDefault >, pcl::DefaultFeatureRepresentation< PointDefault >, pcl::DefaultPointRepresentation< PointDefault >, pcl::DefaultFeatureRepresentation< PFHSignature125 >, pcl::DefaultFeatureRepresentation< PFHRGBSignature250 >, pcl::DefaultFeatureRepresentation< NormalBasedSignature12 >, pcl::DefaultPointRepresentation< Narf36 >, pcl::DefaultFeatureRepresentation< GASDSignature984 >, pcl::DefaultFeatureRepresentation< GASDSignature7992 >, pcl::DefaultFeatureRepresentation< GASDSignature512 >, and pcl::DefaultFeatureRepresentation< FPFHSignature33 >.

Referenced by pcl::PointRepresentation< PointT >::isValid(), and pcl::PointRepresentation< PointT >::vectorize().

◆ getNumberOfDimensions()

template<typename PointT >
int pcl::PointRepresentation< PointT >::getNumberOfDimensions ( ) const
inline

Return the number of dimensions in the point's vector representation.

Definition at line 190 of file point_representation.h.

References pcl::PointRepresentation< PointT >::nr_dimensions_.

◆ isTrivial()

template<typename PointT >
bool pcl::PointRepresentation< PointT >::isTrivial ( ) const
inline

Returns whether this point representation is trivial.

It is trivial if and only if the following conditions hold:

  • the relevant data consists only of float values
  • the vectorize operation directly copies the first nr_dimensions_ elements of PointT to the out array
  • sizeof(PointT) is a multiple of sizeof(float) In short, a trivial point representation converts the input point to a float array that is the same as if the point was reinterpret_casted to a float array of length nr_dimensions_ .

Definition at line 98 of file point_representation.h.

References pcl::PointRepresentation< PointT >::alpha_, and pcl::PointRepresentation< PointT >::trivial_.

◆ isValid()

template<typename PointT >
virtual bool pcl::PointRepresentation< PointT >::isValid ( const PointT p) const
inlinevirtual

Verify that the input point is valid.

Parameters
pThe point to validate

Definition at line 104 of file point_representation.h.

References pcl::PointRepresentation< PointT >::copyToFloatArray(), pcl::PointRepresentation< PointT >::nr_dimensions_, and pcl::PointRepresentation< PointT >::trivial_.

◆ setRescaleValues()

template<typename PointT >
void pcl::PointRepresentation< PointT >::setRescaleValues ( const float *  rescale_array)
inline

Set the rescale values to use when vectorizing points.

Parameters
[in]rescale_arrayThe array/vector of rescale values. Can be of any type that implements the [] operator.

Definition at line 183 of file point_representation.h.

References pcl::PointRepresentation< PointT >::alpha_, and pcl::PointRepresentation< PointT >::nr_dimensions_.

◆ vectorize() [1/3]

template<typename PointT >
void pcl::PointRepresentation< PointT >::vectorize ( const PointT p,
float *  out 
) const
inline

◆ vectorize() [2/3]

template<typename PointT >
template<typename OutputType >
void pcl::PointRepresentation< PointT >::vectorize ( const PointT p,
OutputType &  out 
) const
inline

Convert input point into a vector representation, rescaling by alpha.

Parameters
[in]pthe input point
[out]outThe output vector. Can be of any type that implements the [] operator.

Definition at line 144 of file point_representation.h.

References pcl::PointRepresentation< PointT >::alpha_, pcl::PointRepresentation< PointT >::copyToFloatArray(), and pcl::PointRepresentation< PointT >::nr_dimensions_.

◆ vectorize() [3/3]

template<typename PointT >
void pcl::PointRepresentation< PointT >::vectorize ( const PointT p,
std::vector< float > &  out 
) const
inline

Member Data Documentation

◆ alpha_

template<typename PointT >
std::vector<float> pcl::PointRepresentation< PointT >::alpha_
protected

A vector containing the rescale factor to apply to each dimension.

Definition at line 65 of file point_representation.h.

Referenced by pcl::PointRepresentation< PointT >::isTrivial(), pcl::PointRepresentation< PointT >::setRescaleValues(), and pcl::PointRepresentation< PointT >::vectorize().

◆ nr_dimensions_

template<typename PointT >
int pcl::PointRepresentation< PointT >::nr_dimensions_ = 0
protected

◆ trivial_

template<typename PointT >
bool pcl::PointRepresentation< PointT >::trivial_ = false
protected

Indicates whether this point representation is trivial.

It is trivial if and only if the following conditions hold:

  • the relevant data consists only of float values
  • the vectorize operation directly copies the first nr_dimensions_ elements of PointT to the out array
  • sizeof(PointT) is a multiple of sizeof(float) In short, a trivial point representation converts the input point to a float array that is the same as if the point was reinterpret_casted to a float array of length nr_dimensions_ . This value says that this representation can be trivial; it is only trivial if setRescaleValues() has not been set.

Definition at line 75 of file point_representation.h.

Referenced by pcl::PointRepresentation< PointT >::isTrivial(), and pcl::PointRepresentation< PointT >::isValid().


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