You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

173 lines
6.2 KiB
C++

/*=========================================================================
Program: Visualization Toolkit
Module: vtkPCANormalEstimation.h
Copyright (c) Kitware, Inc.
All rights reserved.
See LICENSE file for details.
This software is distributed WITHOUT ANY WARRANTY; without even
the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE. See the above copyright notice for more information.
=========================================================================*/
/**
* @class vtkPCANormalEstimation
* @brief generate point normals using local tangent planes
*
*
* vtkPCANormalEstimation generates point normals using PCA (principal
* component analysis). Basically this estimates a local tangent plane
* around each sample point p by considering a small neighborhood of points
* around p, and fitting a plane to the neighborhood (via PCA). A good
* introductory reference is Hoppe's "Surface reconstruction from
* unorganized points."
*
* To use this filter, specify a neighborhood size. This may have to be set
* via experimentation. In addition, the user may optionally specify a point
* locator (instead of the default locator), which is used to accelerate
* searches around the sample point. Finally, the user should specify how to
* generate consistently-oriented normals. As computed by PCA, normals may
* point in arbitrary +/- orientation, which may not be consistent with
* neighboring normals. There are three methods to address normal
* consistency: 1) leave the normals as computed, 2) adjust the +/- sign of
* the normals so that the normals all point towards a specified point, and
* 3) perform a traversal of the point cloud and flip neighboring normals so
* that they are mutually consistent.
*
* The output of this filter is the same as the input except that a normal
* per point is produced. (Note that these are unit normals.) While any
* vtkPointSet type can be provided as input, the output is represented by an
* explicit representation of points via a vtkPolyData. This output polydata
* will populate its instance of vtkPoints, but no cells will be defined
* (i.e., no vtkVertex or vtkPolyVertex are contained in the output).
*
* @warning
* This class has been threaded with vtkSMPTools. Using TBB or other
* non-sequential type (set in the CMake variable
* VTK_SMP_IMPLEMENTATION_TYPE) may improve performance significantly.
*
* @sa
* vtkPCACurvatureEstimation
*/
#ifndef vtkPCANormalEstimation_h
#define vtkPCANormalEstimation_h
#include "vtkFiltersPointsModule.h" // For export macro
#include "vtkPolyDataAlgorithm.h"
class vtkAbstractPointLocator;
class vtkIdList;
class VTKFILTERSPOINTS_EXPORT vtkPCANormalEstimation : public vtkPolyDataAlgorithm
{
public:
//@{
/**
* Standard methods for instantiating, obtaining type information, and
* printing information.
*/
static vtkPCANormalEstimation* New();
vtkTypeMacro(vtkPCANormalEstimation, vtkPolyDataAlgorithm);
void PrintSelf(ostream& os, vtkIndent indent) override;
//@}
//@{
/**
* For each sampled point, specify the number of the closest, surrounding
* points used to estimate the normal (the so called k-neighborhood). By
* default 25 points are used. Smaller numbers may speed performance at the
* cost of accuracy.
*/
vtkSetClampMacro(SampleSize, int, 1, VTK_INT_MAX);
vtkGetMacro(SampleSize, int);
//@}
/**
* This enum is used to control how normals oriented is controlled.
*/
enum Style
{
AS_COMPUTED = 0,
POINT = 1,
GRAPH_TRAVERSAL = 3
};
//@{
/**
* Configure how the filter addresses consistency in normal
* oreientation. When initially computed using PCA, a point normal may
* point in the + or - direction, which may not be consistent with
* neighboring points. To address this, various strategies have been used
* to create consistent normals. The simplest approach is to do nothing
* (AsComputed). Another simple approach is to flip the normal based on its
* direction with respect to a specified point (i.e., point normals will
* point towrads the specified point). Finally, a full traversal of points
* across the graph of neighboring, connected points produces the best
* results but is computationally expensive.
*/
vtkSetMacro(NormalOrientation, int);
vtkGetMacro(NormalOrientation, int);
void SetNormalOrientationToAsComputed() { this->SetNormalOrientation(AS_COMPUTED); }
void SetNormalOrientationToPoint() { this->SetNormalOrientation(POINT); }
void SetNormalOrientationToGraphTraversal() { this->SetNormalOrientation(GRAPH_TRAVERSAL); }
//@}
//@{
/**
* If the normal orientation is to be consistent with a specified
* direction, then an orientation point should be set. The sign of the
* normals will be modified so that they point towards this point. By
* default, the specified orientation point is (0,0,0).
*/
vtkSetVector3Macro(OrientationPoint, double);
vtkGetVectorMacro(OrientationPoint, double, 3);
//@}
//@{
/**
* The normal orientation can be flipped by enabling this flag.
*/
vtkSetMacro(FlipNormals, bool);
vtkGetMacro(FlipNormals, bool);
vtkBooleanMacro(FlipNormals, bool);
//@}
//@{
/**
* Specify a point locator. By default a vtkStaticPointLocator is
* used. The locator performs efficient searches to locate points
* around a sample point.
*/
void SetLocator(vtkAbstractPointLocator* locator);
vtkGetObjectMacro(Locator, vtkAbstractPointLocator);
//@}
protected:
vtkPCANormalEstimation();
~vtkPCANormalEstimation() override;
// IVars
int SampleSize;
vtkAbstractPointLocator* Locator;
int NormalOrientation;
double OrientationPoint[3];
bool FlipNormals;
// Methods used to produce consistent normal orientations
void TraverseAndFlip(
vtkPoints* inPts, float* normals, char* pointMap, vtkIdList* wave, vtkIdList* wave2);
// Pipeline management
int RequestData(vtkInformation*, vtkInformationVector**, vtkInformationVector*) override;
int FillInputPortInformation(int port, vtkInformation* info) override;
private:
vtkPCANormalEstimation(const vtkPCANormalEstimation&) = delete;
void operator=(const vtkPCANormalEstimation&) = delete;
};
#endif