swectral.create_example_spec_exp#
- swectral.create_example_spec_exp(dir_path, sample_n=10, n_bands=4, is_regression=True, use_val_group=False)#
Create a standard test SpecExp instance for spectral experiments.
This function generates a SpecExp data manager pre-populated with: :rtype:
SpecExpA test group.
A synthetic multispectral image.
Synthetic ROIs.
Synthetic sample labels and targets.
It can be used for testing spectral processing pipelines, demonstration or development.
- Parameters:
- dir_path
str Directory where the test data (images and ROI XML) will be stored. Will be created if it does not exist.
- sample_n
int,optional Number of samples / ROIs to generate. Default is 10.
- n_bands
int,optional Number of bands in the synthetic image. Default is 4.
- is_regressionbool,
optional If True, target values are numeric (regression). If False, targets are categorical (classification). Default is True.
- use_val_groupbool,
optional Whether validation group is enabled. Default is False.
- dir_path
- Returns:
SpecExpA
SpecExpinstance populated with synthetic data, ready for spectral analysis.
Notes
The synthetic image is a pseudo-random gradient with band-specific variations.
ROIs are rectangular and evenly distributed across the raster height.
Sample labels and target values are automatically generated based on
sample_n.
Examples
Create a default test
SpecExpinstance:>>> exp = create_test_spec_exp("test_spec_exp")
Create a test instance with 20 samples and 6 spectral bands:
>>> exp = create_test_spec_exp("test_spec_exp_dir", sample_n=20, n_bands=6)
Create a classification-type
SpecExpinstance:>>> exp = create_test_spec_exp("class_spec_exp", is_regression=False)