图形处理之曲率计算vtkCurvatures

基本概念

        曲率是曲面弯曲程度的一种度量，是几何体的一种重要的局部特征。如下图所示，要计算曲面上给定点M的曲率，考虑经过M的法线的一个平面与曲面相交，得到一条二维曲线，称之为曲面在M点的一条法截线，如下图中所示的C法；经过M点法向量的曲面可以任意旋转，即可得到任意多条法截线，每条法截线会对应一个曲率，取具有最大曲率和最小曲率的两条法截线作为主法截线，如下图中C1和C2所示，其对应的曲率为k1和k2，称为主曲率；高斯曲率等于主曲率的乘积，即k1xk2；平均曲率等于主曲率k1和k2的平均值，即(k1+k2)/2。以上只是曲率的直观几何解释。

示例演示

        VTK中的vtkCurvatures类实现了四种计算网格模型点曲率的计算方法，四种曲率包括最小曲率、最大曲率、高斯曲率和平均曲率。

curvaturesFilter->SetCurvatureTypeToMinimum();
curvaturesfilter->SetCurvatureTypeToMaximum();
curvaturesFilter->SetCurvatureTypeToGaussian();
curvaturesFilter->SetCurvatureTypeToMean();

        该类接受一个vtkPolyData数据，将计算得到的曲率数据作为网格模型的点的属性数据存入返回的vtkPolyData中。这里简单给出高斯曲率的原理，如下图：

/**********************************************************************

Copyright (c) Mr.Bin. All rights reserved.
For more information visit: http://blog.csdn.net/webzhuce

**********************************************************************/
#include <vtkSmartPointer.h>
#include <vtkCurvatures.h>
#include <vtkPolyDataReader.h>
#include <vtkLookupTable.h>
#include <vtkColorTransferFunction.h>
#include <vtkColorSeries.h>
#include <vtkPointData.h>
#include <vtkPolyDataMapper.h>
#include <vtkActor.h>
#include <vtkScalarBarActor.h>
#include <vtkRenderWindow.h>
#include <vtkRenderer.h>
#include <vtkRenderWindowInteractor.h>

//test data : ../data/fran_cut.vtk
int main(int argc, char *argv[])
{
	vtkNew<vtkPolyDataReader> reader;
	reader->SetFileName("E:/github/Study-VTK/data/fran_cut.vtk");
	reader->Update();

	vtkNew<vtkCurvatures> curvaturesfilter;
	curvaturesfilter->SetInputConnection(reader->GetOutputPort());
	//curvaturesFilter->SetCurvatureTypeToMinimum();
	curvaturesfilter->SetCurvatureTypeToMaximum();
	//curvaturesFilter->SetCurvatureTypeToGaussian();
	//curvaturesFilter->SetCurvatureTypeToMean();
	curvaturesfilter->Update();

	double scalarrange[2];
	curvaturesfilter->GetOutput()->GetScalarRange(scalarrange);
	vtkNew<vtkLookupTable> lut;
	lut->SetHueRange(0.0,0.6);
	lut->SetAlphaRange(1.0,1.0);
	lut->SetValueRange(1.0,1.0);
	lut->SetSaturationRange(1.0,1.0);
	lut->SetNumberOfTableValues(256);
	lut->SetRange(scalarrange);
	lut->Build();

	vtkNew<vtkPolyDataMapper> mapper;
	mapper->SetInputConnection(curvaturesfilter->GetOutputPort());
	mapper->SetLookupTable(lut);
	mapper->SetScalarRange(scalarrange);
	vtkNew<vtkActor> actor;
	actor->SetMapper(mapper);
	vtkNew<vtkScalarBarActor> scalarbar;
	scalarbar->SetLookupTable(mapper->GetLookupTable());
	scalarbar->SetTitle(curvaturesfilter->GetOutput()->GetPointData()->GetScalars()->GetName());
	scalarbar->SetNumberOfLabels(5);
	vtkNew<vtkRenderer> renderer;
	renderer->AddActor(actor);
	renderer->AddActor2D(scalarbar);
	renderer->SetBackground(1.0, 1.0, 1.0);
	vtkNew<vtkRenderWindow> renderwindow;
	renderwindow->AddRenderer(renderer);
	renderwindow->SetSize(640, 480);
	renderwindow->Render();
	renderwindow->SetWindowName("PolyDataCurvature");
	vtkNew<vtkRenderWindowInteractor> renderwindowinteractor;
	renderwindowinteractor->SetRenderWindow(renderwindow);
	renderwindow->Render();
	renderwindowinteractor->Start();

	return EXIT_SUCCESS;
}

运行结果

参考资料

• VTK图形图像开发进阶[M]