Integration of CG and shot footage is tricky: it has to be done carefully and with the end goal clearly defined.
When you shoot your green screen element, add enough markers on it so that you can track it later. Plan the placement of your trackers so that you get a good idea of the actual geometry of the location used. Pay special attention at the places where the tracking marks might overlap with the foreground elements, try to avoid placing markers that are going to be blocked, or that might create probrems with the key.
No you don't need a massive green screen to pull this off, you just need a green screen large enough to cover the edges of the person you want to superimpose on the scene. The rest you can take away with masks (creating what is called a garbage-matte)
Keep accurate notes of what camera was used, the sensor size, lens, camera height, distance to the subject and do include some element on the shot that will help you determine the scale. For example if you set tracking points and measure the distance between them you can later have blender use that scale. To set the scale properly you need to have at least one known distance in the scene.
Do not use a zoom in the shot (as blender cannot deal with variable focal lengths yet). Shoot a distortion chart with the same lens used for the shot to determine lens distortion.
After the shoot bring the elements to blender, track them and reconstruct the scene. and combine the elements using the compositor.
Keep in mind that for a better integration your 3D scene needs to be in a scale that matches the real world.
The whole procedure and workflow is not within the scope of this site but there are plenty of videos out there that can help you. Sebastian Koenig's "Match, Track, Blend 2" series should give you enough information.
Before you do anything please read: How can I get better results when doing camera motion tracking?
If you get stuck please ask specific questions in different posts.