To perform the 3D reconstruction from photographs, ContextCapture must know very accurately the photogroup properties of each input photogroup and the pose of each input photograph. If you ignore these properties, or if you do not know them with sufficient accuracy, ContextCapture can automatically estimate them by a process called aerotriangulation - or aerial triangulation - sometimes abbreviated to AT.

The aerotriangulation starts from an input block, and produces a new block with computed or adjusted properties.

The aerotriangulation can take into account current camera positions (e.g., initialized from GPS), or control points for georeferencing.

Although ContextCapture can compute an aerotriangulation without initial information about the pose of input photographs, it is not recommended for a very large number of photos; in this case the aerotriangulation without any input positioning is unlikely to give satisfying results. Massive datasets should preferably include approximate pose information (e.g. INS) and should be imported from an XML or MS Excel file. ContextCapture can then adjust a very large imported block by aerotriangulation, the only limit in block size being computer memory.