matchTemplate is it appropriate for live camera feed?

The ultimate goal for me is to design a program that in real time uses a live camera feed and can detect billboards with particular images.

Would matchTemplate() be the most appropriate method to use or should I use something else?

know the limitations, it’s neither invariant to rotation, scaling or shearing (all of which you get, if a human points a live camera anywhere)

What is the best/most recommended method(s) I should use?