Question

I am trying to make a program to work on top of an existing GUI to annotate it and provide extra calculations and statistical information. I want to do this using image recognition, as I have learned a fair amount about this in University using Matlab and similar things. I can get a handle to the window I want to perform image recognition on, but I don’t know how to turn that handle into an image of that window, and all its visible child windows. I suppose I am looking for something like the screenshot function, but restricted to a single window.

How would I go about doing this? I suppose I’d need something like a .bmp to mess about with. Also, it would have to be efficient enough that I could call it several times a second without my PC grinding to a halt.

Hopefully this isn’t an obvious question, I typed some things into google but didn’t get anything related.

Answer 1

One simple way is using the PrintWindow API (which is an automated Alt + Print basically). The following example takes a screenshot of the calculator, but you would just have to replace the handles.

void CScreenShotDlg::OnPaint()
{
    // device context for painting
    CPaintDC dc(this);

    // Get the window handle of calculator application.
    HWND hWnd = ::FindWindow( 0, _T( "Calculator" ));

    // Take screenshot.
    PrintWindow( hWnd,
                 dc.GetSafeHdc(),
                 0 );
}

(see https://learn.microsoft.com/en-us/windows/win32/api/winuser/nf-winuser-printwindow )