To build the Structure from Motion Docker images:
OpenMVG + OpenMVS
To process an image set:
./bin/mvg_mvs_pipeline --image-dir <image-dir> --ouput-dir <output-dir> # e.g. ./bin/mvg_mvs_pipeline --image-dir image_sets/kermit --output-dir output/kermit
OpenMVG looks at image EXIF data to discover what camera took the photo. It then consults its own camera database to find the camera's focal length.
If OpenMVG doesn't recognize your camera, you'll need to add your camera's info to sensor_width_camera_database.txt. After, rebuild the OpenMVG Docker image.