Hi,

Maybe this forum is not for this kind of question because it regards Qt.

How is possible to map a current memory area contained by a MatrixXi to a QImage using a QImage::Format_RGB32?

The constructor in QImage accept an unsigned char* as first pointer to the data and then matrix width and height, as well as a format specifier.

Code: Select all

QImage ( uchar * data, int width, int height, Format format );


but the cast from int to unsigned char give me some problem and I really don't know actually if this makes sense.

I would like to show the content of a matrix of integers [0,255] using a QImage on a Qt Widget.

Any hint is appreciated, thanks!

Well, RGB32 uses 3 8-bit values per pixel, one for each channel, whereas your matrix only contains one value per pixel. If you just cast your data, you will probably get senseless results, as it will be 8-bit masks of your values (first 8 bits, second 8 bits, etc.). On the other hand, Qt saves 8-bit images with indexed colors, so it will not be possible to set the values directly. I'd suggest a loop over the pixels setting the values iteratively. It's not as efficient, but probably makes more sense.

Thank you, your suggestion has clarified me the situation and helped me to understand how to do.

To overcome the limitation of obtaining a color image from double data, I wrote a simple map function (which emulates the classical octave jet colormap)

Code: Select all

// Convert a double in [0,1] to a qRbg value
uint MatrixPaintWidget::doubleToJetMap(double x)
{
    double  r = 255*((x >= 0.375 & x < 0.625) * (4.0 * x - 1.5) + (x >= 0.625 & x < 0.875) + (x >= 0.875) * (-4.0 * x + 4.5));
    double g = 255*((x >= 0.125 & x < 0.375) * (4.0 * x - 0.5) + (x >= 0.375 & x < 0.625) + (x >= 0.625 & x < 0.875) * (-4.0 * x + 3.5));
    double b = 255*((x < 0.125) * (4.0 * x + 0.5) + (x >= 0.125 & x < 0.375) + (x >= 0.375 & x < 0.625) * (-4.0 * x + 2.5));

    return qRgb((int)r,(int)g,(int)b);
}


With this map, transforming a MatrixXd to a coloured RGB QImage is quite simple using two for loops.

Code: Select all

QImage eigenToQImage(const Eigen::MatrixXd &m)
{
    QImage I(m.cols(),m.rows(),QImage::Format_RGB32);

    for ( int i=0; i<m.cols(); i++ )
    {
        for ( int j=0; j<m.rows(); j++)
            I.setPixel(j,i, doubleToJetMap(m.coeffRef(i,j)));
    }
    return I;
}


I hope this can be useful to others (important, for the image to extend on the full colormap, matrix values must lie in the interval [0,1]