thinking WAY out of the box...
another way to go would be to allow for inclusion of scaled objects in photos (essentially a scaled image laying against the wall, for example), indexed to known aspects of specific cameras (so you can adjust for known lens lengths et al), that could then do an edge detection and draw a scaled image from a sequence of photos
For instance, starting at the garage and going clockwise around the home, you could quickly generate a shell-then chop it up into GLA and sub areas from that.
I think you'd need to have the camera attached to a pc somehow in order for this to fully work-so it could capture distance and zoom info from the camera...that said, it would make drawings very quick, and allow for easy inclusion of 3-d models if you wanted to, complete with stiched photos and a virtual tour type output. Output it to flash, and make it available to your clients for a small price...could also work well with listing appraisals, as an added feature...
But I'm sure someone's already doing all this somewhere, right?