/* -*- mode: C++ ; c-file-style: "stroustrup" -*- ***************************** * Qwt Widget Library * Copyright (C) 1997 Josef Wilgen * Copyright (C) 2002 Uwe Rathmann * * This library is free software; you can redistribute it and/or * modify it under the terms of the Qwt License, Version 1.0 *****************************************************************************/ #ifndef QWT_CURVE_FITTER_H #define QWT_CURVE_FITTER_H #include "qwt_global.h" #include #include class QwtSpline; /*! \brief Abstract base class for a curve fitter */ class QWT_EXPORT QwtCurveFitter { public: virtual ~QwtCurveFitter(); /*! Find a curve which has the best fit to a series of data points \param polygon Series of data points \return Curve points */ virtual QPolygonF fitCurve( const QPolygonF &polygon ) const = 0; protected: QwtCurveFitter(); private: QwtCurveFitter( const QwtCurveFitter & ); QwtCurveFitter &operator=( const QwtCurveFitter & ); }; /*! \brief A curve fitter using cubic splines */ class QWT_EXPORT QwtSplineCurveFitter: public QwtCurveFitter { public: /*! Spline type The default setting is Auto \sa setFitMode(), FitMode() */ enum FitMode { /*! Use the default spline algorithm for polygons with increasing x values ( p[i-1] < p[i] ), otherwise use a parametric spline algorithm. */ Auto, //! Use a default spline algorithm Spline, //! Use a parametric spline algorithm ParametricSpline }; QwtSplineCurveFitter(); virtual ~QwtSplineCurveFitter(); void setFitMode( FitMode ); FitMode fitMode() const; void setSpline( const QwtSpline& ); const QwtSpline &spline() const; QwtSpline &spline(); void setSplineSize( int ); int splineSize() const; virtual QPolygonF fitCurve( const QPolygonF & ) const; private: QPolygonF fitSpline( const QPolygonF & ) const; QPolygonF fitParametric( const QPolygonF & ) const; class PrivateData; PrivateData *d_data; }; /*! \brief 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. */ class QWT_EXPORT QwtWeedingCurveFitter: public QwtCurveFitter { public: QwtWeedingCurveFitter( double tolerance = 1.0 ); virtual ~QwtWeedingCurveFitter(); void setTolerance( double ); double tolerance() const; void setChunkSize( uint ); uint chunkSize() const; virtual QPolygonF fitCurve( const QPolygonF & ) const; private: virtual QPolygonF simplify( const QPolygonF & ) const; class Line; class PrivateData; PrivateData *d_data; }; #endif