Qwt User's Guide  6.1.2
 All Classes Functions Variables Typedefs Enumerations Enumerator Pages
Public Member Functions | List of all members
QwtWeedingCurveFitter Class Reference

A curve fitter implementing Douglas and Peucker algorithm. More...

#include <qwt_curve_fitter.h>

Inheritance diagram for QwtWeedingCurveFitter:
Inheritance graph
[legend]

Public Member Functions

 QwtWeedingCurveFitter (double tolerance=1.0)
 
virtual ~QwtWeedingCurveFitter ()
 Destructor.
 
void setTolerance (double)
 
double tolerance () const
 
void setChunkSize (uint)
 
uint chunkSize () const
 
virtual QPolygonF fitCurve (const QPolygonF &) const
 
- Public Member Functions inherited from QwtCurveFitter
virtual ~QwtCurveFitter ()
 Destructor.
 

Additional Inherited Members

- Protected Member Functions inherited from QwtCurveFitter
 QwtCurveFitter ()
 Constructor.
 

Detailed Description

A curve fitter implementing Douglas and Peucker algorithm.

The purpose of the Douglas and Peucker algorithm is that given a 'curve' composed of line segments to find a curve not too dissimilar but that has fewer points. The algorithm defines 'too dissimilar' based on the maximum distance (tolerance) between the original curve and the smoothed curve.

The runtime of the algorithm increases non linear ( worst case O( n*n ) ) and might be very slow for huge polygons. To avoid performance issues it might be useful to split the polygon ( setChunkSize() ) and to run the algorithm for these smaller parts. The disadvantage of having no interpolation at the borders is for most use cases irrelevant.

The smoothed curve consists of a subset of the points that defined the original curve.

In opposite to QwtSplineCurveFitter the Douglas and Peucker algorithm reduces the number of points. By adjusting the tolerance parameter according to the axis scales QwtSplineCurveFitter can be used to implement different level of details to speed up painting of curves of many points.

Constructor & Destructor Documentation

QwtWeedingCurveFitter::QwtWeedingCurveFitter ( double  tolerance = 1.0)

Constructor

Parameters
toleranceTolerance
See Also
setTolerance(), tolerance()

Member Function Documentation

uint QwtWeedingCurveFitter::chunkSize ( ) const
Returns
Maximum for the number of points passed to a run of the algorithm - or 0, when unlimited
See Also
setChunkSize()
QPolygonF QwtWeedingCurveFitter::fitCurve ( const QPolygonF &  points) const
virtual
Parameters
pointsSeries of data points
Returns
Curve points

Implements QwtCurveFitter.

void QwtWeedingCurveFitter::setChunkSize ( uint  numPoints)

Limit the number of points passed to a run of the algorithm

The runtime of the Douglas Peucker algorithm increases non linear with the number of points. For a chunk size > 0 the polygon is split into pieces passed to the algorithm one by one.

Parameters
numPointsMaximum for the number of points passed to the algorithm
See Also
chunkSize()
void QwtWeedingCurveFitter::setTolerance ( double  tolerance)

Assign the tolerance

The tolerance is the maximum distance, that is acceptable between the original curve and the smoothed curve.

Increasing the tolerance will reduce the number of the resulting points.

Parameters
toleranceTolerance
See Also
tolerance()
double QwtWeedingCurveFitter::tolerance ( ) const
Returns
Tolerance
See Also
setTolerance()