Detection of a very large object (vehicle) in a single static camera

I’m trying to make a deep learning model for an automated detection and classification of vehicles (i.e cars, bus, trucks, etc) in a paid parking lot, but since the cameras are too up close for a large vehicle to fit in a single frame, it is creating an issue.

I thought of extracting the moving vehicle(s) from the video and classify it in real-time using image stitching technique but I’m kinda new to this image stitching thing so any help will be appreciated.

Input will be a video file from a single static camera and vehicles may or may not fit in single frame. (Example: the cameras you see in a toll booth or a parking lot)