2.2. Dense image matching
The development of 3D scanners over the last fifteen years
might have been expected to lead to a loss of interest in photogrammetry
and its applications. In fact, the universal availability of
powerful computers has led to renewed interest in digital photogrammetry,
and dense image matching algorithms for producing
automatic 3D reconstructions are still being developed (Hartley and
Zisserman, 2004; Ahmadabadian et al., 2013).
Matching of stereoscopic pairs is based on the identification and
mapping of tie points. Tie-point detection is mostly based on
analyzing the radiometry of individual pixels and their immediate
vicinities, and may be guided by geometric properties such as the
epipolar constraint (Hartley and Zisserman, 2004). This Dense
Stereo Matching (DStM) process increases computing speeds and
enables the optimization of results. Point cloud density is chosen by
the operator during the matching operation. The minimum