Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
233 lines (202 sloc) 9.76 KB
Program: Visualization Toolkit
Module: vtkImageToPolyDataFilter.h
Copyright (c) Ken Martin, Will Schroeder, Bill Lorensen
All rights reserved.
See Copyright.txt or for details.
This software is distributed WITHOUT ANY WARRANTY; without even
PURPOSE. See the above copyright notice for more information.
// .NAME vtkImageToPolyDataFilter - generate linear primitives (vtkPolyData) from an image
// .SECTION Description
// vtkImageToPolyDataFilter converts raster data (i.e., an image) into
// polygonal data (i.e., quads or n-sided polygons), with each polygon
// assigned a constant color. This is useful for writers that generate vector
// formats (i.e., CGM or PostScript). To use this filter, you specify how to
// quantize the color (or whether to use an image with a lookup table), and
// what style the output should be. The output is always polygons, but the
// choice is n x m quads (where n and m define the input image dimensions)
// "Pixelize" option; arbitrary polygons "Polygonalize" option; or variable
// number of quads of constant color generated along scan lines "RunLength"
// option.
// The algorithm quantizes color in order to create coherent regions that the
// polygons can represent with good compression. By default, the input image
// is quantized to 256 colors using a 3-3-2 bits for red-green-blue. However,
// you can also supply a single component image and a lookup table, with the
// single component assumed to be an index into the table. (Note: a quantized
// image can be generated with the filter vtkImageQuantizeRGBToIndex.) The
// number of colors on output is equal to the number of colors in the input
// lookup table (or 256 if the built in linear ramp is used).
// The output of the filter is polygons with a single color per polygon cell.
// If the output style is set to "Polygonalize", the polygons may have an
// large number of points (bounded by something like 2*(n+m)); and the
// polygon may not be convex which may cause rendering problems on some
// systems (use vtkTriangleFilter). Otherwise, each polygon will have four
// vertices. The output also contains scalar data defining RGB color in
// unsigned char form.
// .SECTION Caveats
// The input linear lookup table must
// be of the form of 3-component unsigned char.
// This filter defines constant cell colors. If you have a plotting
// device that supports Gouraud shading (linear interpolation of color), then
// superior algorithms are available for generating polygons from images.
// Note that many plotting devices/formats support only a limited number of
// colors.
// .SECTION See Also
// vtkCGMWriter vtkImageQuantizeRGBToIndex vtkTriangleFilter
#ifndef __vtkImageToPolyDataFilter_h
#define __vtkImageToPolyDataFilter_h
#include "vtkPolyDataAlgorithm.h"
class vtkAppendPolyData;
class vtkDataArray;
class vtkEdgeTable;
class vtkIdTypeArray;
class vtkIntArray;
class vtkScalarsToColors;
class vtkStructuredPoints;
class vtkTimeStamp;
class vtkUnsignedCharArray;
class VTK_HYBRID_EXPORT vtkImageToPolyDataFilter : public vtkPolyDataAlgorithm
void PrintSelf(ostream& os, vtkIndent indent);
// Description:
// Instantiate object with initial number of colors 256.
static vtkImageToPolyDataFilter* New();
// Description:
// Specify how to create the output. Pixelize means converting the image
// to quad polygons with a constant color per quad. Polygonalize means
// merging colors together into polygonal regions, and then smoothing
// the regions (if smoothing is turned on). RunLength means creating
// quad polygons that may encompass several pixels on a scan line. The
// default behavior is Polygonalize.
void SetOutputStyleToPixelize()
void SetOutputStyleToPolygonalize()
void SetOutputStyleToRunLength()
// Description:
// Specify how to quantize color.
void SetColorModeToLUT()
void SetColorModeToLinear256()
// Description:
// Set/Get the vtkLookupTable to use. The lookup table is used when the
// color mode is set to LUT and a single component scalar is input.
virtual void SetLookupTable(vtkScalarsToColors*);
// Description:
// If the output style is set to polygonalize, then you can control
// whether to smooth boundaries.
vtkSetMacro(Smoothing, int);
vtkGetMacro(Smoothing, int);
vtkBooleanMacro(Smoothing, int);
// Description:
// Specify the number of smoothing iterations to smooth polygons. (Only
// in effect if output style is Polygonalize and smoothing is on.)
// Description:
// Turn on/off whether the final polygons should be decimated.
// whether to smooth boundaries.
vtkSetMacro(Decimation, int);
vtkGetMacro(Decimation, int);
vtkBooleanMacro(Decimation, int);
// Description:
// Specify the error to use for decimation (if decimation is on).
// The error is an absolute number--the image spacing and
// dimensions are used to create points so the error should be
// consistent with the image size.
// Description:
// Specify the error value between two colors where the colors are
// considered the same. Only use this if the color mode uses the
// default 256 table.
// Description:
// Specify the size (n by n pixels) of the largest region to
// polygonalize. When the OutputStyle is set to VTK_STYLE_POLYGONALIZE,
// large amounts of memory are used. In order to process large images,
// the image is broken into pieces that are at most Size pixels in
// width and height.
virtual int RequestData(vtkInformation *, vtkInformationVector **, vtkInformationVector *);
virtual int FillInputPortInformation(int port, vtkInformation *info);
int OutputStyle;
int ColorMode;
int Smoothing;
int NumberOfSmoothingIterations;
int Decimation;
double DecimationError;
int Error;
int SubImageSize;
vtkScalarsToColors *LookupTable;
virtual void PixelizeImage(vtkUnsignedCharArray *pixels, int dims[3],
double origin[3], double spacing[3],
vtkPolyData *output);
virtual void PolygonalizeImage(vtkUnsignedCharArray *pixels, int dims[3],
double origin[3], double spacing[3],
vtkPolyData *output);
virtual void RunLengthImage(vtkUnsignedCharArray *pixels, int dims[3],
double origin[3], double spacing[3],
vtkPolyData *output);
vtkUnsignedCharArray *Table; // color table used to quantize points
vtkTimeStamp TableMTime;
int *Visited; // traverse & mark connected regions
vtkUnsignedCharArray *PolyColors; // the colors of each region -> polygon
vtkEdgeTable *EdgeTable; // keep track of intersection points
vtkEdgeTable *EdgeUseTable; // keep track of polygons use of edges
vtkIntArray *EdgeUses; //the two polygons that use an edge
//and point id associated with edge (if any)
vtkAppendPolyData *Append;
void BuildTable(unsigned char *inPixels);
vtkUnsignedCharArray *QuantizeImage(vtkDataArray *inScalars, int numComp,
int type, int dims[3], int ext[4]);
int ProcessImage(vtkUnsignedCharArray *pixels, int dims[2]);
int BuildEdges(vtkUnsignedCharArray *pixels, int dims[3], double origin[3],
double spacing[3], vtkUnsignedCharArray *pointDescr,
vtkPolyData *edges);
void BuildPolygons(vtkUnsignedCharArray *pointDescr, vtkPolyData *edges,
int numPolys, vtkUnsignedCharArray *polyColors);
void SmoothEdges(vtkUnsignedCharArray *pointDescr, vtkPolyData *edges);
void DecimateEdges(vtkPolyData *edges, vtkUnsignedCharArray *pointDescr,
double tol2);
void GeneratePolygons(vtkPolyData *edges, int numPolys, vtkPolyData *output,
vtkUnsignedCharArray *polyColors,
vtkUnsignedCharArray *pointDescr);
int GetNeighbors(unsigned char *ptr, int &i, int &j, int dims[3],
unsigned char *neighbors[4], int mode);
void GetIJ(int id, int &i, int &j, int dims[3]);
unsigned char *GetColor(unsigned char *rgb);
int IsSameColor(unsigned char *p1, unsigned char *p2);
vtkImageToPolyDataFilter(const vtkImageToPolyDataFilter&); // Not implemented.
void operator=(const vtkImageToPolyDataFilter&); // Not implemented.