There are many areas where highly overlapped imagery are being developed through UAS as well as terresterial methods such as vehicle mounted 'ladybug' cameras and others that have georeferenced photo centers. Feature extraction from imagery has been realized for more than 20 years and more recently 3D items can be created from 2D imagery. If one considers the output of traditional aerial imagery products then the availability to pull features from imagery has been around for more that 50 years.
So my idea is this: Create an extension that can extract 3d positions from a collection of 2D pictures.
