Tools for analyzing and georeferencing 3D models. The necessary files are obtained with an openMVG pipeline and stored on the 'gargantua' share.
This python script can be used to transform the coordinates of a model from its arbitrary referential to a chosen cartesian geographical referential. The file containing the homologous points, necessary for computing a rigid transform from one 3D referential to the other, is of the following format:
source_point_1_x source_point_1_y source_point_1_z target_point_1_x target_point_1_y target_point_1_z
source_point_2_x source_point_2_y source_point_2_z target_point_2_x target_point_2_y target_point_2_z
...
Ideally the target points are selected on a cadaster map of the chosen referential.
python vodom_to_csv.py --model <path-to-the-model-to-be-georeferenced>
--points <path-to-the-homologous-points>
--epsg <epsg-code-of-the-chosen-referential>
When a model is computed, it often happens that all input images do not appear on the final model. To get the list of the names of missing images, a simple python script can be used as following:
python get_missing_files.py --json <path-to-the-sfm-data-file>