The reference point is wherever you want it to be. If you define it
as the tip of the nose of one of the people, that's fine. Alternately,
it could be one of the claws on the dog across the street, or a leaf
on the tree that's directly over the people.
Since you're planning on describing the motion of the car, you
probably don't want to pick the car itself as the reference point.
If you did that, then the car would never move, but everything else
would.