def process(self, inputs, outputs):
# Benchmarking
start_time = system.now_raw()
log.ODM_INFO('Running ODM OpenSfM Cell')
# get inputs
tree = inputs.tree
args = inputs.args
reconstruction = inputs.reconstruction
photos = reconstruction.photos
if not photos:
log.ODM_ERROR('Not enough photos in photos array to start OpenSfM')
return ecto.QUIT
# create working directories
system.mkdir_p(tree.opensfm)
# check if we rerun cell or not
rerun_cell = (args.rerun is not None and
args.rerun == 'opensfm') or \
(args.rerun_all) or \
(args.rerun_from is not None and
'opensfm' in args.rerun_from)
if args.fast_orthophoto:
output_file = io.join_paths(tree.opensfm, 'reconstruction.ply')
elif args.use_opensfm_dense:
output_file = tree.opensfm_model
else:
output_file = tree.opensfm_reconstruction
# check if reconstruction was done before
if not io.file_exists(output_file) or rerun_cell:
# create file list
list_path = io.join_paths(tree.opensfm, 'image_list.txt')
has_alt = True
with open(list_path, 'w') as fout:
for photo in photos:
if not photo.altitude:
has_alt = False
fout.write('%s\n' % io.join_paths(tree.dataset_raw, photo.filename))
# create config file for OpenSfM
config = [
"use_exif_size: %s" % ('no' if not self.params.use_exif_size else 'yes'),
"feature_process_size: %s" % self.params.feature_process_size,
"feature_min_frames: %s" % self.params.feature_min_frames,
"processes: %s" % self.params.processes,
"matching_gps_neighbors: %s" % self.params.matching_gps_neighbors,
"depthmap_method: %s" % args.opensfm_depthmap_method,
"depthmap_resolution: %s" % args.depthmap_resolution,
"depthmap_min_patch_sd: %s" % args.opensfm_depthmap_min_patch_sd,
"depthmap_min_consistent_views: %s" % args.opensfm_depthmap_min_consistent_views,
"optimize_camera_parameters: %s" % ('no' if self.params.fixed_camera_params else 'yes')
]
if has_alt:
log.ODM_DEBUG("Altitude data detected, enabling it for GPS alignment")
config.append("use_altitude_tag: yes")
config.append("align_method: naive")
else:
config.append("align_method: orientation_prior")
config.append("align_orientation_prior: vertical")
if args.use_hybrid_bundle_adjustment:
log.ODM_DEBUG("Enabling hybrid bundle adjustment")
config.append("bundle_interval: 100") # Bundle after adding 'bundle_interval' cameras
config.append("bundle_new_points_ratio: 1.2") # Bundle when (new points) / (bundled points) > bundle_new_points_ratio
config.append("local_bundle_radius: 1") # Max image graph distance for images to be included in local bundle adjustment
if args.matcher_distance > 0:
config.append("matching_gps_distance: %s" % self.params.matching_gps_distance)
if tree.odm_georeferencing_gcp:
config.append("bundle_use_gcp: yes")
io.copy(tree.odm_georeferencing_gcp, tree.opensfm)
# write config file
log.ODM_DEBUG(config)
config_filename = io.join_paths(tree.opensfm, 'config.yaml')
with open(config_filename, 'w') as fout:
fout.write("\n".join(config))
# run OpenSfM reconstruction
matched_done_file = io.join_paths(tree.opensfm, 'matching_done.txt')
if not io.file_exists(matched_done_file) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm extract_metadata %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
system.run('PYTHONPATH=%s %s/bin/opensfm detect_features %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
system.run('PYTHONPATH=%s %s/bin/opensfm match_features %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
with open(matched_done_file, 'w') as fout:
fout.write("Matching done!\n")
else:
log.ODM_WARNING('Found a feature matching done progress file in: %s' %
matched_done_file)
if not io.file_exists(tree.opensfm_tracks) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm create_tracks %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
else:
log.ODM_WARNING('Found a valid OpenSfM tracks file in: %s' %
tree.opensfm_tracks)
if not io.file_exists(tree.opensfm_reconstruction) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm reconstruct %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
else:
log.ODM_WARNING('Found a valid OpenSfM reconstruction file in: %s' %
tree.opensfm_reconstruction)
# Check that a reconstruction file has been created
if not io.file_exists(tree.opensfm_reconstruction):
log.ODM_ERROR("The program could not process this dataset using the current settings. "
"Check that the images have enough overlap, "
"that there are enough recognizable features "
"and that the images are in focus. "
"You could also try to increase the --min-num-features parameter."
"The program will now exit.")
sys.exit(1)
# Always export VisualSFM's reconstruction and undistort images
# as we'll use these for texturing (after GSD estimation and resizing)
if not args.ignore_gsd:
image_scale = gsd.image_scale_factor(args.orthophoto_resolution, tree.opensfm_reconstruction)
else:
image_scale = 1.0
if not io.file_exists(tree.opensfm_reconstruction_nvm) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm export_visualsfm --image_extension png --scale_focal %s %s' %
(context.pyopencv_path, context.opensfm_path, image_scale, tree.opensfm))
else:
log.ODM_WARNING('Found a valid OpenSfM NVM reconstruction file in: %s' %
tree.opensfm_reconstruction_nvm)
# These will be used for texturing
system.run('PYTHONPATH=%s %s/bin/opensfm undistort --image_format png --image_scale %s %s' %
(context.pyopencv_path, context.opensfm_path, image_scale, tree.opensfm))
# Skip dense reconstruction if necessary and export
# sparse reconstruction instead
if args.fast_orthophoto:
system.run('PYTHONPATH=%s %s/bin/opensfm export_ply --no-cameras %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
elif args.use_opensfm_dense:
# Undistort images at full scale in JPG
# (TODO: we could compare the size of the PNGs if they are < than depthmap_resolution
# and use those instead of re-exporting full resolution JPGs)
system.run('PYTHONPATH=%s %s/bin/opensfm undistort %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
system.run('PYTHONPATH=%s %s/bin/opensfm compute_depthmaps %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
else:
log.ODM_WARNING('Found a valid OpenSfM reconstruction file in: %s' %
tree.opensfm_reconstruction)
# check if reconstruction was exported to bundler before
if not io.file_exists(tree.opensfm_bundle_list) or rerun_cell:
# convert back to bundler's format
system.run('PYTHONPATH=%s %s/bin/export_bundler %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
else:
log.ODM_WARNING('Found a valid Bundler file in: %s' %
tree.opensfm_reconstruction)
if reconstruction.georef:
system.run('PYTHONPATH=%s %s/bin/opensfm export_geocoords %s --transformation --proj \'%s\'' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm, reconstruction.georef.projection.srs))
outputs.reconstruction = reconstruction
if args.time:
system.benchmark(start_time, tree.benchmarking, 'OpenSfM')
log.ODM_INFO('Running ODM OpenSfM Cell - Finished')
return ecto.OK if args.end_with != 'opensfm' else ecto.QUIT
def process(self, inputs, outputs):
# Benchmarking
start_time = system.now_raw()
log.ODM_INFO('Running ODM OpenSfM Cell')
# get inputs
tree = inputs.tree
args = inputs.args
reconstruction = inputs.reconstruction
photos = reconstruction.photos
if not photos:
log.ODM_ERROR('Not enough photos in photos array to start OpenSfM')
return ecto.QUIT
# create working directories
system.mkdir_p(tree.opensfm)
# check if we rerun cell or not
rerun_cell = (args.rerun is not None and
args.rerun == 'opensfm') or \
(args.rerun_all) or \
(args.rerun_from is not None and
'opensfm' in args.rerun_from)
if args.fast_orthophoto:
output_file = io.join_paths(tree.opensfm, 'reconstruction.ply')
elif args.use_opensfm_dense:
output_file = tree.opensfm_model
else:
output_file = tree.opensfm_reconstruction
# check if reconstruction was done before
if not io.file_exists(output_file) or rerun_cell:
# create file list
list_path = io.join_paths(tree.opensfm, 'image_list.txt')
has_alt = True
with open(list_path, 'w') as fout:
for photo in photos:
if not photo.altitude:
has_alt = False
fout.write('%s\n' %
io.join_paths(tree.dataset_raw, photo.filename))
# create config file for OpenSfM
config = [
"use_exif_size: %s" %
('no' if not self.params.use_exif_size else 'yes'),
"feature_process_size: %s" % self.params.feature_process_size,
"feature_min_frames: %s" % self.params.feature_min_frames,
"processes: %s" % self.params.processes,
"matching_gps_neighbors: %s" %
self.params.matching_gps_neighbors,
"depthmap_method: %s" % args.opensfm_depthmap_method,
"depthmap_resolution: %s" % args.depthmap_resolution,
"depthmap_min_patch_sd: %s" %
args.opensfm_depthmap_min_patch_sd,
"depthmap_min_consistent_views: %s" %
args.opensfm_depthmap_min_consistent_views,
"optimize_camera_parameters: %s" %
('no' if self.params.fixed_camera_params else 'yes')
]
if has_alt:
log.ODM_DEBUG(
"Altitude data detected, enabling it for GPS alignment")
config.append("use_altitude_tag: yes")
config.append("align_method: naive")
else:
config.append("align_method: orientation_prior")
config.append("align_orientation_prior: vertical")
if args.use_hybrid_bundle_adjustment:
log.ODM_DEBUG("Enabling hybrid bundle adjustment")
config.append(
"bundle_interval: 100"
) # Bundle after adding 'bundle_interval' cameras
config.append(
"bundle_new_points_ratio: 1.2"
) # Bundle when (new points) / (bundled points) > bundle_new_points_ratio
config.append(
"local_bundle_radius: 1"
) # Max image graph distance for images to be included in local bundle adjustment
if args.matcher_distance > 0:
config.append("matching_gps_distance: %s" %
self.params.matching_gps_distance)
if tree.odm_georeferencing_gcp:
config.append("bundle_use_gcp: yes")
io.copy(tree.odm_georeferencing_gcp, tree.opensfm)
# write config file
log.ODM_DEBUG(config)
config_filename = io.join_paths(tree.opensfm, 'config.yaml')
with open(config_filename, 'w') as fout:
fout.write("\n".join(config))
# run OpenSfM reconstruction
matched_done_file = io.join_paths(tree.opensfm,
'matching_done.txt')
if not io.file_exists(matched_done_file) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm extract_metadata %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
system.run('PYTHONPATH=%s %s/bin/opensfm detect_features %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
system.run('PYTHONPATH=%s %s/bin/opensfm match_features %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
with open(matched_done_file, 'w') as fout:
fout.write("Matching done!\n")
else:
log.ODM_WARNING(
'Found a feature matching done progress file in: %s' %
matched_done_file)
if not io.file_exists(tree.opensfm_tracks) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm create_tracks %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
else:
log.ODM_WARNING('Found a valid OpenSfM tracks file in: %s' %
tree.opensfm_tracks)
if not io.file_exists(tree.opensfm_reconstruction) or rerun_cell:
system.run('PYTHONPATH=%s %s/bin/opensfm reconstruct %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
else:
log.ODM_WARNING(
'Found a valid OpenSfM reconstruction file in: %s' %
tree.opensfm_reconstruction)
# Check that a reconstruction file has been created
if not io.file_exists(tree.opensfm_reconstruction):
log.ODM_ERROR(
"The program could not process this dataset using the current settings. "
"Check that the images have enough overlap, "
"that there are enough recognizable features "
"and that the images are in focus. "
"You could also try to increase the --min-num-features parameter."
"The program will now exit.")
sys.exit(1)
# Always export VisualSFM's reconstruction and undistort images
# as we'll use these for texturing (after GSD estimation and resizing)
if not args.ignore_gsd:
image_scale = gsd.image_scale_factor(
args.orthophoto_resolution, tree.opensfm_reconstruction)
else:
image_scale = 1.0
if not io.file_exists(
tree.opensfm_reconstruction_nvm) or rerun_cell:
system.run(
'PYTHONPATH=%s %s/bin/opensfm export_visualsfm --image_extension png --scale_focal %s %s'
% (context.pyopencv_path, context.opensfm_path,
image_scale, tree.opensfm))
else:
log.ODM_WARNING(
'Found a valid OpenSfM NVM reconstruction file in: %s' %
tree.opensfm_reconstruction_nvm)
# These will be used for texturing
system.run(
'PYTHONPATH=%s %s/bin/opensfm undistort --image_format png --image_scale %s %s'
% (context.pyopencv_path, context.opensfm_path, image_scale,
tree.opensfm))
# Skip dense reconstruction if necessary and export
# sparse reconstruction instead
if args.fast_orthophoto:
system.run(
'PYTHONPATH=%s %s/bin/opensfm export_ply --no-cameras %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
elif args.use_opensfm_dense:
# Undistort images at full scale in JPG
# (TODO: we could compare the size of the PNGs if they are < than depthmap_resolution
# and use those instead of re-exporting full resolution JPGs)
system.run('PYTHONPATH=%s %s/bin/opensfm undistort %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
system.run(
'PYTHONPATH=%s %s/bin/opensfm compute_depthmaps %s' %
(context.pyopencv_path, context.opensfm_path,
tree.opensfm))
else:
log.ODM_WARNING(
'Found a valid OpenSfM reconstruction file in: %s' %
tree.opensfm_reconstruction)
# check if reconstruction was exported to bundler before
if not io.file_exists(tree.opensfm_bundle_list) or rerun_cell:
# convert back to bundler's format
system.run(
'PYTHONPATH=%s %s/bin/export_bundler %s' %
(context.pyopencv_path, context.opensfm_path, tree.opensfm))
else:
log.ODM_WARNING('Found a valid Bundler file in: %s' %
tree.opensfm_reconstruction)
if reconstruction.georef:
system.run(
'PYTHONPATH=%s %s/bin/opensfm export_geocoords %s --transformation --proj \'%s\''
% (context.pyopencv_path, context.opensfm_path, tree.opensfm,
reconstruction.georef.projection.srs))
outputs.reconstruction = reconstruction
if args.time:
system.benchmark(start_time, tree.benchmarking, 'OpenSfM')
log.ODM_INFO('Running ODM OpenSfM Cell - Finished')
return ecto.OK if args.end_with != 'opensfm' else ecto.QUIT
def process(self, args, outputs):
tree = outputs['tree']
reconstruction = outputs['reconstruction']
photos = reconstruction.photos
if not photos:
log.ODM_ERROR('Not enough photos in photos array to start OpenSfM')
exit(1)
octx = OSFMContext(tree.opensfm)
octx.setup(args,
tree.dataset_raw,
photos,
gcp_path=tree.odm_georeferencing_gcp,
rerun=self.rerun())
octx.extract_metadata(self.rerun())
self.update_progress(20)
octx.feature_matching(self.rerun())
self.update_progress(30)
octx.reconstruct(self.rerun())
self.update_progress(70)
# If we find a special flag file for split/merge we stop right here
if os.path.exists(octx.path("split_merge_stop_at_reconstruction.txt")):
log.ODM_INFO("Stopping OpenSfM early because we found: %s" %
octx.path("split_merge_stop_at_reconstruction.txt"))
self.next_stage = None
return
if args.fast_orthophoto:
output_file = octx.path('reconstruction.ply')
elif args.use_opensfm_dense:
output_file = tree.opensfm_model
else:
output_file = tree.opensfm_reconstruction
# Always export VisualSFM's reconstruction and undistort images
# as we'll use these for texturing (after GSD estimation and resizing)
if not args.ignore_gsd:
image_scale = gsd.image_scale_factor(args.orthophoto_resolution,
tree.opensfm_reconstruction)
else:
image_scale = 1.0
if not io.file_exists(tree.opensfm_reconstruction_nvm) or self.rerun():
octx.run(
'export_visualsfm --image_extension png --scale_focal %s' %
image_scale)
else:
log.ODM_WARNING(
'Found a valid OpenSfM NVM reconstruction file in: %s' %
tree.opensfm_reconstruction_nvm)
# These will be used for texturing
undistorted_images_path = octx.path("undistorted")
if not io.dir_exists(undistorted_images_path) or self.rerun():
octx.run('undistort --image_format png --image_scale %s' %
image_scale)
else:
log.ODM_WARNING("Found an undistorted directory in %s" %
undistorted_images_path)
self.update_progress(80)
# Skip dense reconstruction if necessary and export
# sparse reconstruction instead
if args.fast_orthophoto:
if not io.file_exists(output_file) or self.rerun():
octx.run('export_ply --no-cameras')
else:
log.ODM_WARNING("Found a valid PLY reconstruction in %s" %
output_file)
elif args.use_opensfm_dense:
# Undistort images at full scale in JPG
# (TODO: we could compare the size of the PNGs if they are < than depthmap_resolution
# and use those instead of re-exporting full resolution JPGs)
if not io.file_exists(output_file) or self.rerun():
octx.run('undistort')
octx.run('compute_depthmaps')
else:
log.ODM_WARNING("Found a valid dense reconstruction in %s" %
output_file)
# check if reconstruction was exported to bundler before
octx.export_bundler(tree.opensfm_bundle_list, self.rerun())
self.update_progress(90)
if reconstruction.georef and (not io.file_exists(
tree.opensfm_transformation) or self.rerun()):
octx.run('export_geocoords --transformation --proj \'%s\'' %
reconstruction.georef.projection.srs)
else:
log.ODM_WARNING("Will skip exporting %s" %
tree.opensfm_transformation)