eradiate.experiments.CanopyAtmosphereExperiment#

class eradiate.experiments.CanopyAtmosphereExperiment(measures=_Nothing.NOTHING, illumination=_Nothing.NOTHING, geometry='plane_parallel', atmosphere=_Nothing.NOTHING, canopy=None, padding=0, surface=_Nothing.NOTHING, integrator=AUTO)[source]#

Bases: eradiate.experiments._core.EarthObservationExperiment

Simulate radiation in a scene with an explicit canopy and atmosphere. This experiment assumes that the surface is plane and accounts for ground unit cell padding.

Parameters

measures (list of Measure or list of dict or Measure or dict, default: MultiDistantMeasure()) – List of measure specifications. The passed list may contain dictionaries, which will be interpreted by measure_factory. Optionally, a single Measure or dictionary specification may be passed and will automatically be wrapped into a list.
illumination (DirectionalIllumination or ConstantIllumination or dict, default: DirectionalIllumination()) – Illumination specification. This parameter can be specified as a dictionary which will be interpreted by illumination_factory.
geometry (str or dict or AtmosphereGeometry, default: plane_parallel) – Atmosphere geometry.
atmosphere (Atmosphere or dict or None, optional, default: HomogeneousAtmosphere()) – Atmosphere specification. If set to None, no atmosphere will be added. This parameter can be specified as a dictionary which will be interpreted by atmosphere_factory.
canopy (Canopy or dict or None, optional, default: None) – Canopy specification. This parameter can be specified as a dictionary which will be interpreted by biosphere_factory.
padding (int, optional, default: 0) – Padding level. The canopy will be padded with copies to account for adjacency effects. This, in practice, has effects similar to making the scene periodic.A value of 0 will yield only the defined scene. A value of 1 will add one copy in every direction, yielding a 3×3 patch. A value of 2 will yield a 5×5 patch, etc. The optimal padding level depends on the scene.
surface (Surface or BSDF or dict or None, optional, default: BasicSurface(bsdf=LambertianBSDF())) – Surface specification. A Surface object may be passed: its shape specifications will be bypassed and the surface size will be computed automatically upon kernel dictionary generation. A BSDF may also be passed: it will be wrapped automatically in a BasicSurface instance. If a dictionary is passed, it will be first interpreted as a BSDF; if this fails, it will then be interpreted as a Surface. Finally, this field can be set to None: in that case, no surface will be added.
integrator (Integrator or dict or AUTO, default: AUTO) – Monte Carlo integration algorithm specification. This parameter can be specified as a dictionary which will be interpreted by IntegratorFactory.convert().If set to AUTO, the integrator will be set depending on the presence of an atmosphere.If an atmosphere is defined the integrator defaults to VolPathIntegrator`otherwise a :class:.PathIntegrator` will be used.

Fields

measures (list of Measure) – List of measure specifications.
illumination (DirectionalIllumination or ConstantIllumination) – Illumination specification.
geometry (PlaneParallelGeometry or SphericalShellGeometry) – Atmosphere geometry.
atmosphere (Atmosphere or None) – Atmosphere specification.
canopy (Canopy or None) – Canopy specification.
padding (int) – Padding level.
surface (Surface or None) – Surface specification.

Warning

Canopy padding is controlled using the padding parameter: do not pad the canopy itself manually.

Notes

A post-initialisation step will constrain the measure setup if a distant measure is used and no target is defined:
- if a canopy is defined, the target will be set to the top of the canopy unit cell (i.e. without its padding);
- if no canopy is defined, the target will be set according to the atmosphere (i.e. to [0, 0, toa] where toa is the top-of-atmosphere altitude);
- if neither atmosphere nor canopy are defined, the target is set to [0, 0, 0].
This experiment supports arbitrary measure positioning, except for MultiRadiancemeterMeasure, for which subsensor origins are required to be either all inside or all outside of the atmosphere. If an unsuitable configuration is detected, a ValueError will be raised during initialisation.

classmethod from_dict(d)#

Instantiate from a dictionary. The default implementation raises an exception.

Parameters: d (dict) – Dictionary to be converted to an Experiment.
Returns: Experiment

kernel_dict(ctx)[source]#

Return a dictionary suitable for kernel scene configuration.

Parameters: ctx (KernelDictContext) – A context data structure containing parameters relevant for kernel dictionary generation.
Returns: KernelDict – Kernel dictionary which can be loaded as a Mitsuba object.

kernel_dicts(measure)#

A generator which returns kernel dictionaries (and the associated context) relevant to a given measure.

Parameters

measure (Measure or int) – Measure for which kernel dictionaries are to be generated. Alternatively, the index in the self.measure list can be passed.

Yields

kernel_dict (KernelDict) – Generated kernel dictionary.
ctx (KernelDictContext) – Context used to generate kernel_dict.

pipeline(*measures)#

Request post-processing pipeline for a given measure.

Parameters: *measures (Measure or int) – One or several measures for which to get a post-processing pipeline. If integer values are passed, they are used to query the measure list.
Returns: pipelines (Pipeline or tuple of Pipeline) – If a single measure is passed, a single Pipeline instance is returned; if multiple measures are passed, a tuple of pipelines is returned.

postprocess(*measures, pipeline_kwargs=None)#

Post-process raw results stored in a measure’s results field. This requires a successful execution of process(). Post-processed results are stored in self.results.

Parameters

*measures (Measure or int) – One or several measures for which to perform post-processing. Alternatively, indexes in the measure array can be passed. If no value is passed, all measures are processed.
pipeline_kwargs (dict, optional) – A dictionary of pipeline keyword arguments forwarded to Pipeline.transform().

Raises

ValueError – If measure.results is None, i.e. if process() has not been successfully run.