Need help with my C++ and the SDK

stuw

Hi all,

I'm playing with the SDK on Linux, but not having much luck. I want to try working with QT's own graphics libraries rather than OpenCV which is making it a little harder (especially as I'm coming from over a decade of windows VB programming).

So far I can open the camera and get/set settings, but when trying to get a frame with getImageData I think i'm filling the frame variable, but I'm having problems passing it to QImage. No matter what I do, I cannot get the frame data into an image.

This is what I have so far.


bool MainWindow:()
{
    QMessageBox msgBox;
    int CamNum = ui->cbCameras->currentIndex();
    bool bresult = openCamera(CamNum);
    if(!bresult)
    {
        msgBox.setText("Error connecting to camera" + QString:(CamNum));
        msgBox.setStandardButtons(QMessageBox:);
        msgBox.setDefaultButton(QMessageBox:);
        msgBox.exec();
        return false;
    }

    initCamera();
    ui->btnConnectCamera->setText("Disconnect");
    this->setWindowTitle("Sensor Temp: " + QString:(getSensorTemp()));
    setImageFormat(320,240,1,IMG_Y8);
    setValue(CONTROL_EXPOSURE, 33*1000, true);
    startCapture();
    int bufferSize = getWidth()*getHeight()*8;
    unsigned char* cFrame;
    cFrame = (unsigned char*)malloc(bufferSize);
    while (true) {
        getImageData((unsigned char*) cFrame, bufferSize, -1);
        QImage image = QImage:((unsigned char*)cFrame,bufferSize,0);
        ui->lblImage->setPixmap(QPixmap:(image));
        ui->lblImage->show();
    }
    return true;
}

Any tips qould be appreciated..

Sam@ZWO

Hi Stuart ,I am not familar with QT.

but can you get data from the function getImageData?

stuw

Yes, I think so. Looking at the memory address of the return variable it gets filled, but for some reason the QT image handling routines don't recognise it.

jamesf

Just seen this post, Stu.

I think the problem here is that you're trying to use the wrong method for creating the QImage. I believe that QImage: is supposed to be for creating a QImage from some other recognised file format such as PNG or JPEG. If you're loading it from a raw video frame you can't do it that way because you just have the raw data -- no header or any other such information to allow QImage to work out what type the frame is.

What I've done is something like:


newImage = new QImage (( const uchar* ) imageData, xSize, ySize, QImage: );

This is for a colour frame, not a greyscale Y8 format. Unfortunately there doesn't appear to be a way to create a QImage straight from a greyscale format. Probably it would require that the greyscale data be duplicated for all three colour planes and then use RGB888.

I should admit however that I'm still having problems getting colour images from the SDK myself. The header files suggest that it's possible to get a debayered 8bpp RGB image, but when I try to display it (using exactly the same code that displays a debayered 8bbp RGB image from my SPC900), it isn't working properly.

James