python.workflows.clfSPTSetup Namespace Reference

Analyses a set of labelled point cloud tiles and sets up a new SPT project directory. Running clfSPTSetup is the mandatory first step in the training workflow before proceeding with clfSPTPreprocess and clfSPTTrain. More...

Classes
class	clfSPTSetup

Functions
tuple	scan_dataset (list files, logger)
	Scans all input files. More...
dict	build_class_map (dict found_ids, list ignore_ids, dict overrides)
	Builds the class mapping {odm_id -> name}.
None	print_class_preview (dict found_ids, dict class_map, list ignore_ids)
	Tabular display of the detected classes.
Optional[dict]	interactive_confirm (dict found_ids, dict class_map, list ignore_ids)
tuple	detect_features (set available_attrs, logger)
	Determines from the available ODM attributes which SPT features can be used. More...
dict	create_split (list files, dict per_file_counts, str split_type, dict ratios, logger)
None	report_split_composition (dict per_file_counts, dict split_result, logger)
dict	create_project_structure (str project_dir, logger)
	Creates the project directories and returns their paths.
None	copy_split_files (list files, dict split_result, dict paths, logger)
None	write_dataset_config_py (str output_path, str datensatz, dict class_map, list ignore_ids, dict feature_flags, dict split_result)
	Generates {dataset}_config.py.
None	write_cfg_yaml (str output_path, str datensatz, str project_dir, list point_hf, dict feat_size)
	Generates {dataset}.cfg as a YAML file. More...

Variables
	message
	SPT_ROOT = os.path.join(_DISTRO_ROOT, 'addons', 'superpoint_transformer', 'superpoint_transformer')
	SPT_CONFIG_DIR = os.path.join(SPT_ROOT, 'configs')
dictionary	ASPRS_CLASSES
list	DEFAULT_IGNORE_IDS = [0, 1, 7, 18]
dictionary	ATTRIBUTE_TO_FEATURE
list	GEOMETRIC_FEATURES
	script

Detailed Description

Analyses a set of labelled point cloud tiles and sets up a new SPT project directory. Running clfSPTSetup is the mandatory first step in the training workflow before proceeding with clfSPTPreprocess and clfSPTTrain.

This script serves as the entry point for the SPT training workflow. Its primary job is to generate the necessary configuration files and directory structure required by the downstream tools. To do this, it scans the point cloud files (supporting all formats handled by pyDM.Import) to determine their semantic class distributions and check which point attributes are available for model training. Based on the class distributions, it creates either a stratified or random train/validation/test split and organises everything into a clean project structure containing <project>.cfg and <project>_config.py. For adjusting the dataset splits it is not necessary to re-run the script, it is also possible to move files between the split folders manually.

See also

Script documentation

Function Documentation

copy_split_files()

None python.workflows.clfSPTSetup.copy_split_files	(	list	files,
		dict	split_result,
		dict	paths,
			logger
	)

Copies the tiles into their corresponding split directories.

create_split()

dict python.workflows.clfSPTSetup.create_split	(	list	files,
		dict	per_file_counts,
		str	split_type,
		dict	ratios,
			logger
	)

Creates the train/val/test split.
Returns {split_name: [file_stems]}.

detect_features()

tuple python.workflows.clfSPTSetup.detect_features	(	set	available_attrs,
			logger
	)

Determines from the available ODM attributes which SPT features can be used.

Returns (point_hf_list, feat_size_dict, feature_flags_dict).

interactive_confirm()

Optional[dict] python.workflows.clfSPTSetup.interactive_confirm	(	dict	found_ids,
		dict	class_map,
		list	ignore_ids
	)

Displays the class table and allows interactive changes.
Returns: updated class_map, or None (cancelled).

References python.workflows.clfSPTSetup.print_class_preview().

report_split_composition()

None python.workflows.clfSPTSetup.report_split_composition	(	dict	per_file_counts,
		dict	split_result,
			logger
	)

Checks the class distribution per split and warns if a class is
COMPLETELY missing from a split (typically: a rare class like Building
drops out of test). Without this warning, metrics for the missing
class in that split are silently meaningless.

scan_dataset()

tuple python.workflows.clfSPTSetup.scan_dataset	(	list	files,
			logger
	)

Scans all input files.

Returns (total_class_counts, available_attributes, per_file_counts).

write_cfg_yaml()

None python.workflows.clfSPTSetup.write_cfg_yaml	(	str	output_path,
		str	datensatz,
		str	project_dir,
		list	point_hf,
		dict	feat_size
	)

Generates {dataset}.cfg as a YAML file.

Contains all model parameters for clfSPTTrain.py. Can be edited manually before training.

Variable Documentation

ASPRS_CLASSES

dictionary ASPRS_CLASSES

Initial value:

=  {
    0:  'Never Classified',
    1:  'Unclassified',
    2:  'Ground',
    3:  'Low Vegetation',
    4:  'Medium Vegetation',
    5:  'High Vegetation',
    6:  'Building',
    7:  'Low Point (Noise)',
    8:  'Reserved',
    9:  'Water',
    10: 'Rail',
    11: 'Road Surface',
    12: 'Reserved (Overlap)',
    13: 'Wire - Guard',
    14: 'Wire - Conductor',
    15: 'Transmission Tower',
    16: 'Wire-Structure Connector',
    17: 'Bridge Deck',
    18: 'High Noise',
}

ATTRIBUTE_TO_FEATURE

dictionary ATTRIBUTE_TO_FEATURE

Initial value:

=  {
    'Amplitude':    ('intensity',   1),
    '_Reflectance': ('reflectance', 1),
    'EchoRatio':    ('echo_ratio',  1),
    'NrOfEchos':    ('num_returns', 1)
}

GEOMETRIC_FEATURES

list GEOMETRIC_FEATURES

Initial value:

=  [
    'linearity', 'planarity', 'scattering', 'verticality', 'elevation',
]