Several algorithms to compute point-clouds geometric characteristics (curvature, density, etc.) More...

#include <GeometricalAnalysisTools.h>

Inheritance diagram for CCCoreLib::GeometricalAnalysisTools:

Collaboration diagram for CCCoreLib::GeometricalAnalysisTools:

Public Types
enum	GeomCharacteristic { Feature , Curvature , LocalDensity , ApproxLocalDensity , Roughness , MomentOrder1 }

enum	Density { DENSITY_KNN = 1 , DENSITY_2D , DENSITY_3D }
	Density measurement. More...

enum	ErrorCode { NoError = 0 , InvalidInput = -1 , NotEnoughPoints = -2 , OctreeComputationFailed = -3 , ProcessFailed = -4 , UnhandledCharacteristic = -5 , NotEnoughMemory = -6 , ProcessCancelledByUser = -7 }

Static Public Member Functions
static ErrorCode	ComputeCharactersitic (GeomCharacteristic c, int subOption, GenericIndexedCloudPersist cloud, PointCoordinateType kernelRadius, const CCVector3 roughnessUpDir=nullptr, GenericProgressCallback progressCb=nullptr, DgmOctree inputOctree=nullptr)
	Unified way to compute a geometric characteristic. More...

static ErrorCode	ComputeLocalDensityApprox (GenericIndexedCloudPersist cloud, Density densityType, GenericProgressCallback progressCb=nullptr, DgmOctree *inputOctree=nullptr)
	Computes the local density (approximate) More...

static CCVector3	ComputeGravityCenter (GenericCloud *theCloud)
	Computes the gravity center of a point cloud. More...

static CCVector3	ComputeWeightedGravityCenter (GenericCloud theCloud, ScalarField weights)
	Computes the weighted gravity center of a point cloud. More...

static SquareMatrixd	ComputeCrossCovarianceMatrix (GenericCloud P, GenericCloud Q, const CCVector3 &pGravityCenter, const CCVector3 &qGravityCenter)
	Computes the cross covariance matrix between two clouds (same size) More...

static SquareMatrixd	ComputeWeightedCrossCovarianceMatrix (GenericCloud P, GenericCloud Q, const CCVector3 &pGravityCenter, const CCVector3 &qGravityCenter, ScalarField *coupleWeights=nullptr)
	Computes the cross covariance matrix between two clouds (same size) - weighted version. More...

static SquareMatrixd	ComputeCovarianceMatrix (GenericCloud theCloud, const PointCoordinateType _gravityCenter=nullptr)
	Computes the covariance matrix of a clouds. More...

static ErrorCode	FlagDuplicatePoints (GenericIndexedCloudPersist theCloud, double minDistanceBetweenPoints=std::numeric_limits< double >::epsilon(), GenericProgressCallback progressCb=nullptr, DgmOctree *inputOctree=nullptr)
	Flag duplicate points. More...

static ErrorCode	DetectSphereRobust (GenericIndexedCloudPersist cloud, double outliersRatio, CCVector3 &center, PointCoordinateType &radius, double &rms, GenericProgressCallback progressCb=nullptr, double confidence=0.99, unsigned seed=0)
	Tries to detect a sphere in a point cloud. More...

static ErrorCode	ComputeSphereFrom4 (const CCVector3 &A, const CCVector3 &B, const CCVector3 &C, const CCVector3 &D, CCVector3 &center, PointCoordinateType &radius)
	Computes the center and radius of a sphere passing through 4 points. More...

Static Protected Member Functions
static bool	ComputeGeomCharacteristicAtLevel (const DgmOctree::octreeCell &cell, void *additionalParameters, NormalizedProgress nProgress=nullptr)
	Computes geom characteristic inside a cell. More...

static bool	ComputeApproxPointsDensityInACellAtLevel (const DgmOctree::octreeCell &cell, void *additionalParameters, NormalizedProgress nProgress=nullptr)
	Computes approximate point density inside a cell. More...

static bool	FlagDuplicatePointsInACellAtLevel (const DgmOctree::octreeCell &cell, void *additionalParameters, NormalizedProgress nProgress=nullptr)
	Flags duplicate points inside a cell. More...

static bool	RefineSphereLS (GenericIndexedCloudPersist *cloud, CCVector3 &center, PointCoordinateType &radius, double minReltaiveCenterShift=1.0e-3)
	Refines the estimation of a sphere by (iterative) least-squares.

Detailed Description

Several algorithms to compute point-clouds geometric characteristics (curvature, density, etc.)

Member Enumeration Documentation

◆ Density

enum CCCoreLib::GeometricalAnalysisTools::Density

Density measurement.

Enumerator
DENSITY_KNN	The number of points inside the neighborhing sphere
DENSITY_2D	The number of points divided by the area of the circle that has the same radius as the neighborhing sphere (2D approximation)
DENSITY_3D	The number of points divided by the neighborhing sphere volume (3D)

◆ GeomCharacteristic

enum CCCoreLib::GeometricalAnalysisTools::GeomCharacteristic

Enumerator
Feature	See Neighbourhood::GeomFeature
Curvature	See Neighbourhood::CurvatureType
LocalDensity	Accurate local density (see GeometricalAnalysisTools::Density)
ApproxLocalDensity	Approximate local density (see GeometricalAnalysisTools::Density)
Roughness	Roughness
MomentOrder1	1st order moment

Member Function Documentation

◆ ComputeApproxPointsDensityInACellAtLevel()

bool GeometricalAnalysisTools::ComputeApproxPointsDensityInACellAtLevel	(	const DgmOctree::octreeCell &	cell,
		void **	additionalParameters,
		NormalizedProgress *	nProgress = `nullptr`
	)

staticprotected

Computes approximate point density inside a cell.

Parameters

cell	structure describing the cell on which processing is applied
additionalParameters	see method description
nProgress	optional (normalized) progress notification (per-point)

◆ ComputeCharactersitic()

GeometricalAnalysisTools::ErrorCode GeometricalAnalysisTools::ComputeCharactersitic	(	GeomCharacteristic	c,
		int	subOption,
		GenericIndexedCloudPersist *	cloud,
		PointCoordinateType	kernelRadius,
		const CCVector3 *	roughnessUpDir = `nullptr`,
		GenericProgressCallback *	progressCb = `nullptr`,
		DgmOctree *	inputOctree = `nullptr`
	)

static

Unified way to compute a geometric characteristic.

Once the main geometric characterstic is chosen, the subOption parameter is used to specify the actual feature / curvature type / local density computation algorithm if necessary.

Parameters

c	geometric characterstic
subOption	feature / curvature type / local density computation algorithm or nothing (0)
cloud	cloud to compute the characteristic on
kernelRadius	neighbouring sphere radius
roughnessUpDir	up direction to compute signed roughness values (optional)
progressCb	client application can get some notification of the process progress through this callback mechanism (see GenericProgressCallback)
inputOctree	if not set as input, octree will be automatically computed.

Returns: succes

◆ ComputeCovarianceMatrix()

SquareMatrixd GeometricalAnalysisTools::ComputeCovarianceMatrix	(	GenericCloud *	theCloud,
		const PointCoordinateType *	_gravityCenter = `nullptr`
	)

static

Computes the covariance matrix of a clouds.

Warning: this method uses the cloud global iterator

Parameters

theCloud	point cloud
_gravityCenter	if available, its gravity center

Returns: covariance matrix

◆ ComputeCrossCovarianceMatrix()

SquareMatrixd GeometricalAnalysisTools::ComputeCrossCovarianceMatrix	(	GenericCloud *	P,
		GenericCloud *	Q,
		const CCVector3 &	pGravityCenter,
		const CCVector3 &	qGravityCenter
	)

static

Computes the cross covariance matrix between two clouds (same size)

Used in the ICP algorithm between the cloud to register and the "Closest Points Set" determined from the reference cloud.

Warning: this method uses the clouds global iterators

Parameters

P	the cloud to register
Q	the "Closest Point Set"
pGravityCenter	the gravity center of P
qGravityCenter	the gravity center of Q

Returns: cross covariance matrix

◆ ComputeGeomCharacteristicAtLevel()

bool GeometricalAnalysisTools::ComputeGeomCharacteristicAtLevel	(	const DgmOctree::octreeCell &	cell,
		void **	additionalParameters,
		NormalizedProgress *	nProgress = `nullptr`
	)

staticprotected

Computes geom characteristic inside a cell.

Parameters

cell	structure describing the cell on which processing is applied
additionalParameters	see method description
nProgress	optional (normalized) progress notification (per-point)

◆ ComputeGravityCenter()

CCVector3 GeometricalAnalysisTools::ComputeGravityCenter ( GenericCloud * theCloud )

static

Computes the gravity center of a point cloud.

Warning: this method uses the cloud global iterator

Parameters

theCloud cloud

Returns: gravity center

◆ ComputeLocalDensityApprox()

GeometricalAnalysisTools::ErrorCode GeometricalAnalysisTools::ComputeLocalDensityApprox	(	GenericIndexedCloudPersist *	cloud,
		Density	densityType,
		GenericProgressCallback *	progressCb = `nullptr`,
		DgmOctree *	inputOctree = `nullptr`
	)

static

Computes the local density (approximate)

Old method (based only on the distance to the nearest neighbor).

Warning: As only one neighbor is extracted, the DENSITY_KNN type corresponds in fact to the (inverse) distance to the nearest neighbor.; This method assumes the input scalar field is different from the output one.

Parameters

cloud	processed cloud
densityType	the 'type' of density to compute
progressCb	client application can get some notification of the process progress through this callback mechanism (see GenericProgressCallback)
inputOctree	if not set as input, octree will be automatically computed.

Returns: success (0) or error code (<0)

◆ ComputeSphereFrom4()

GeometricalAnalysisTools::ErrorCode GeometricalAnalysisTools::ComputeSphereFrom4	(	const CCVector3 &	A,
		const CCVector3 &	B,
		const CCVector3 &	C,
		const CCVector3 &	D,
		CCVector3 &	center,
		PointCoordinateType &	radius
	)

static

Computes the center and radius of a sphere passing through 4 points.

Parameters

[in]	A	first point
[in]	B	second point
[in]	C	third point
[in]	D	fourth point
[out]	center	center of the sphere
[out]	radius	radius of the sphere

Returns: success

◆ ComputeWeightedCrossCovarianceMatrix()

SquareMatrixd GeometricalAnalysisTools::ComputeWeightedCrossCovarianceMatrix	(	GenericCloud *	P,
		GenericCloud *	Q,
		const CCVector3 &	pGravityCenter,
		const CCVector3 &	qGravityCenter,
		ScalarField *	coupleWeights = `nullptr`
	)

static

Computes the cross covariance matrix between two clouds (same size) - weighted version.

Used in the ICP algorithm between the cloud to register and the "Closest Points Set" determined from the reference cloud.

Warning: this method uses the clouds global iterators

Parameters

P	the cloud to register
Q	the "Closest Point Set"
pGravityCenter	the gravity center of P
qGravityCenter	the gravity center of Q
coupleWeights	weights for each (Pi,Qi) couple (optional)

Returns: weighted cross covariance matrix

◆ ComputeWeightedGravityCenter()

CCVector3 GeometricalAnalysisTools::ComputeWeightedGravityCenter	(	GenericCloud *	theCloud,
		ScalarField *	weights
	)

static

Computes the weighted gravity center of a point cloud.

Warning: this method uses the cloud global iterator

Parameters

theCloud	cloud
weights	per point weights (only absolute values are considered)

Returns: gravity center

◆ DetectSphereRobust()

GeometricalAnalysisTools::ErrorCode GeometricalAnalysisTools::DetectSphereRobust	(	GenericIndexedCloudPersist *	cloud,
		double	outliersRatio,
		CCVector3 &	center,
		PointCoordinateType &	radius,
		double &	rms,
		GenericProgressCallback *	progressCb = `nullptr`,
		double	confidence = `0.99`,
		unsigned	seed = `0`
	)

static

Tries to detect a sphere in a point cloud.

Inspired from "Parameter Estimation Techniques: A Tutorial with Application to Conic Fitting" by Zhengyou Zhang (Inria Technical Report 2676). More specifically the section 9.5 about Least Median of Squares.

Parameters

[in]	cloud	input cloud
[in]	outliersRatio	proportion of outliers (between 0 and 1)
[out]	center	center of the detected sphere
[out]	radius	radius of the detected sphere
[out]	rms	residuals RMS for the detected sphere
[in]	progressCb	for progress notification (optional)
[in]	confidence	probability that the detected sphere is the right one (strictly below 1)
[in]	seed	if different than 0, this seed will be used for random numbers generation (instead of a random one)

Returns: success

◆ FlagDuplicatePoints()

GeometricalAnalysisTools::ErrorCode GeometricalAnalysisTools::FlagDuplicatePoints	(	GenericIndexedCloudPersist *	theCloud,
		double	minDistanceBetweenPoints = `std::numeric_limits<double>::epsilon()`,
		GenericProgressCallback *	progressCb = `nullptr`,
		DgmOctree *	inputOctree = `nullptr`
	)

static

Flag duplicate points.

This method only requires an output scalar field. Duplicate points will be associated to scalar value 1 (and 0 for the others).

Parameters

theCloud	processed cloud
minDistanceBetweenPoints	min distance between (output) points
progressCb	client application can get some notification of the process progress through this callback mechanism (see GenericProgressCallback)
inputOctree	if not set as input, octree will be automatically computed.

Returns: success (0) or error code (<0)

◆ FlagDuplicatePointsInACellAtLevel()

bool GeometricalAnalysisTools::FlagDuplicatePointsInACellAtLevel	(	const DgmOctree::octreeCell &	cell,
		void **	additionalParameters,
		NormalizedProgress *	nProgress = `nullptr`
	)

staticprotected

Flags duplicate points inside a cell.

Parameters

cell	structure describing the cell on which processing is applied
additionalParameters	see method description
nProgress	optional (normalized) progress notification (per-point)

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

GeometricalAnalysisTools.h
GeometricalAnalysisTools.cpp

Public Types

Static Public Member Functions

Static Protected Member Functions

Detailed Description

Member Enumeration Documentation

◆ Density

◆ GeomCharacteristic

Member Function Documentation

◆ ComputeApproxPointsDensityInACellAtLevel()

◆ ComputeCharactersitic()

◆ ComputeCovarianceMatrix()

◆ ComputeCrossCovarianceMatrix()

◆ ComputeGeomCharacteristicAtLevel()

◆ ComputeGravityCenter()

◆ ComputeLocalDensityApprox()

◆ ComputeSphereFrom4()

◆ ComputeWeightedCrossCovarianceMatrix()

◆ ComputeWeightedGravityCenter()

◆ DetectSphereRobust()

◆ FlagDuplicatePoints()

◆ FlagDuplicatePointsInACellAtLevel()