To sum it up, me and a few other team members are trying to do the same thing FF7 did. When cloud and all the characters walked around the environment, the surroundings were nothing more than a 2D rendered image, but there was also invisible barriers and other things that kept the characters uniform to the scene. So I'm wondering, how do we go about that? We can render our scene images in blender but how do we actually stick animated character models over a rendered image and keep it consistent with the scene? Like say foe example, the character gets further away from the camera and gets smaller, the character climbs the image of a rendered ladder and goes upward and performs a climbing animation. Stuff like that. We greatly appreciate any help, thanks.

As far as i understand you could search for: pre-rendered graphics based engine.
One way i imagine you would do this is to keep also, besides the pre-rendered images a low-poly model of the world used for collisions and physics. This world will also be used to integrate 3D elements with the pre-rendered image. For this to work you need to reconstruct the depth buffer of the pre-rendered image:
position your camera at the exact point and with the exact parameters as when the image was rendered,disable color writes and render.
Now you have a depth buffer and you can go on an render other 3D objects and get correct integration with the pre-rendered scene.