Cross hairs are meant for aiming (centering) and are not usually interactive.
The rectangular visualization's primary purpose is not centering of focus, but selecting the relevant part of the image for scanning. Theoretically user can change both size and location in the screen, therefore this is not a common cross hair.
The rectangular visualization enables the user to see and modify the "region of interest". (See definition of "region of interest" or ROI here [Wikipedia], [Mathworks], [Wolfram]).
I found that this rectange is refered to as "viewfinder box", "viewfinder rectange", "viewfinder area", "crosshair box" (or "annoying yellow box around crosshair" in non-QR usage), however a more correct term (which is commonly used in image processing) would be "region of interest".