Afraid you'll have to raise a ticket with Mendix to find that out.
I believe it simulates a simple download command given by your browser with the image address as a query parameter (similar to how <a href="/images/myw3schoolsimage.jpg" download> would work).
Hope this helps
Mendix works with an Apache framework called PDFBox which in-turn works with Apache FoP.
The document template you create in mendix will be the XSL file and the object that you pass to the document will be an XML file. Because FoP to generate a Pdf document requires XSL and XML file basically.
Then FoP takes care of generating the document and save it as a FileDocument in Mendix, which is then a simple download action.
Hope this clarifies a bit more.