QwtPlotSpectrogram with log scales

I have 2d histogram data that I want to display. This essentially requires that I can set an arbitrary scale to match the binning used to calculate the data set with QwtMatrixRasterData::setValueMatrix. I posted about this on SO. I also managed to make this work the way that I want. I posted this answer that explains my solution.

Uwe suggested that I usw a QwtRasterData based solution but my previous experiments had shows that this does not lead to the result that I want as it re-scales the image to the scales of the plot.

As I have a working solution, I guess my only questions are:

a) is this the right way to approach the situation?
b) does a different solution exists that works OOTB?
c) is there some other way to get Qwt to draw a density plot so that I don't have to do the binning myself?

Code:

---

#include <QApplication>
#include <QMainWindow>
#include <QRect>
#include <QStyleFactory>
 
#include <qwt_color_map.h>
#include <qwt_matrix_raster_data.h>
#include <qwt_plot.h>
#include <qwt_plot_curve.h>
#include <qwt_plot_spectrogram.h>
#include <qwt_scale_engine.h>
 
class PlotSpectrogram : public QwtPlotSpectrogram {
public:
 
  void draw(
      QPainter* painter,
      const QwtScaleMap& xMap,
      const QwtScaleMap& yMap,
      const QRectF& canvasRect ) const override {
 
    QwtScaleMap xMapLin( xMap );
    QwtScaleMap yMapLin( yMap );
 
    auto const xi = data()->interval( Qt::XAxis );
    auto const yi = data()->interval( Qt::YAxis );
 
    auto const dx = xMapLin.transform( xMap.s1() );
    xMapLin.setScaleInterval( xi.minValue(), xi.maxValue() );
    auto const dy = yMapLin.transform( yMap.s2() );
    yMapLin.setScaleInterval( yi.minValue(), yi.maxValue() );
 
    xMapLin.setTransformation( new QwtNullTransform() );
    yMapLin.setTransformation( new QwtNullTransform() );
 
    QwtPlotSpectrogram::draw(
        painter, xMapLin, yMapLin, canvasRect.translated( dx, -dy ) );
  }
};
 
int
main( int argc, char* argv[] ) {
 
  QApplication app( argc, argv );
  QMainWindow wnd;
 
  QVector<double> heat_values( 100 * 100 );
  for( int n = 0; n < 100 * 100; ++n ) {
    heat_values[n] = ( n % 100 ) + n / 100;
  };
 
  //QwtPlotSpectrogram heat;
  PlotSpectrogram heat;
  auto heat_data = std::make_unique<QwtMatrixRasterData>();
  heat_data->setValueMatrix( heat_values, 100 );
  heat_data->setInterval( Qt::XAxis, QwtInterval( 0, 100.0 ) );
  heat_data->setInterval( Qt::YAxis, QwtInterval( 0, 100.0 ) );
  heat_data->setInterval( Qt::ZAxis, QwtInterval( 0, 200.0 ) );
 
  heat.setDisplayMode( QwtPlotSpectrogram::DisplayMode::ImageMode, true );
  heat.setColorMap( new QwtLinearColorMap( Qt::white, Qt::black ) );
  heat.setData( heat_data.release() );
 
  QwtPlot p;
  p.setAxisScaleEngine( QwtPlot::yLeft, new QwtLogScaleEngine() );
  p.setAxisScale( QwtPlot::yLeft, 20.0, 50.0 );
  p.setAxisScaleEngine( QwtPlot::xBottom, new QwtLogScaleEngine() );
  p.setAxisScale( QwtPlot::xBottom, 200, 500.0 );
  p.setAutoDelete( false );
  heat.attach( &p );
  p.repaint();
 
  wnd.setCentralWidget( &p );
  wnd.resize( 400, 300 );
  wnd.show();
 
  return QApplication::exec();
}

---