pylawr.transform.filter

pylawr.transform.filter.clutterfilter.ClutterFilter(...)	Abstract base class for clutter filters
pylawr.transform.filter.ClutterMap([name, ...])	Class to transform/filter the RadarField on basis of a clutter map, calculated with calc_cluttermap of ClutterFilter
pylawr.transform.filter.RINGFilter([...])	Calculates gradients of consecutive range gates for clutter detection.
pylawr.transform.filter.SNR([threshold])	Calculates the Signal-to-Noise ratio.
pylawr.transform.filter.SpeckleFilter([...])	The speckle filter checks the number of rain pixels in a neighborhood.
pylawr.transform.filter.SPKFilter([...])	Calculates gradients of consecutive radar beams for clutter detection.
pylawr.transform.filter.SPINFilter([...])	Calculate SPIN change of the reflectivity for clutter detection according to Hubbert et al. (2009).
pylawr.transform.filter.TDBZFilter([...])	Calculate texture of reflectivity (TDBZ) for clutter detection according to Hubbert et al. (2009).
pylawr.transform.filter.TemporalFilter([...])	The temporal clutter filter compares current radar image with n-saved radar images.

pylawr.transform.filter.clutterfilter.ClutterFilter

class pylawr.transform.filter.clutterfilter.ClutterFilter(threshold)

Bases: ABC

Abstract base class for clutter filters

threshold

The threshold of the reflectivities to filter (dBZ)

Type

float

_abc_impl = <_abc._abc_data object>

abstract calc_map(refl)

Computes field for detecting clutter based on the filter.

Parameters

refl (xarray.DataArray) – The reflectivity array to operate on.

abstract create_cluttermap(refl=None)

Creates clutter map based on the field and threshold.

Returns

• cmap (ClutterMap) – The cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

• refl (xarray.DataArray) – The reflectivity array to operate on.

pylawr.transform.filter.ClutterMap

class pylawr.transform.filter.ClutterMap(name=None, cmap=None, fuzzy_threshold=1)

Bases: Transformer

Class to transform/filter the RadarField on basis of a clutter map, calculated with calc_cluttermap of ClutterFilter

Parameters

• name (string) – The name of used clutter filter.

• cmap (numpy.ndarray) – Probability of clutter for each grid point.

• fuzzy_threshold (float) – Relative value to the number of clutter maps, which indicate clutter, to fulfill clutter condition. The interval of this fuzzy threshold is [0, 1]. A value of 1 describes that every cluttermap has to detect clutter so that this value is filtered out.

layers

Dictionary with zero to many clutter maps

Type

collections.OrderedDict

weights

Dictionary with the weights to corresponding clutter maps.

Type

collections.OrderedDict

fuzzy_threshold

Relative value to the number of clutter maps, which indicate clutter, to fulfill clutter condition. The interval of this fuzzy threshold is [0, 1]. A value of 1 describes that every cluttermap has to detect clutter so that this value is filtered out.

Type

float

_abc_impl = <_abc._abc_data object>

append(other_clutter)

Appends an other instance of pylawr.filter.cluttermap.ClutterMap to this instance.

Parameters

other_clutter (pylawr.filter.cluttermap.ClutterMap) – Some instance of this class. The layers of this class are used to update the layers of this cluttermap.

property array

Get the values of the cluttermaps as array.

Returns

layer_array – The concatenated values of the different cluttermaps in preserved order. The cluttermaps are concatenated over a new first dimension.

Return type

numpy.ndarray

mean()

Get the weighted mean of this cluttermap. The cluttermap is weighted with set weights.

Returns

mean_values – The weighted mean values with the same shape as a single cluttermap.

Return type

py:class:numpy.ndarray

transform(array, *args, **kwargs)

Transform a given array with a polar grid according to the ClutterMap object. The clutter maps in the layers attribute are applied on the radar field. If the number of clutter maps are greater or equal to the fuzzy threshold at a specific grid point, clutter is detected.

Parameters

array (xarray.DataArray) – The reflectivity array to operate on.

Returns

transformed_array – The array with removed clutter as NaN-values

Return type

xarray.DataArray

pylawr.transform.filter.RINGFilter

class pylawr.transform.filter.RINGFilter(threshold=3.0, window_size=11, window_criterion=0.5, ring_width=1)

Bases: ClutterFilter

Calculates gradients of consecutive range gates for clutter detection. If the difference between consecutive range gates exceeds a threshold for percentage (window_criterion) of consecutive beams within a window, the gate is identified as clutter. This filter is similar to the SPKFilter, only the axis is changed.

With the default parameters a ring of six consecutive beams is identified as clutter.

threshold

The threshold of the reflectivities to filter (dBZ)

Type

float

_wz

The number of consecutive beams in which the threshold has to be exceeded

Type

int

_wc

The percentage of consecutive beam gates the criteria have to be fulfilled

Type

float

_rw

The assumed ring width to detect.

Type

int

_abc_impl = <_abc._abc_data object>

calc_map(array)

Computes ring field for detecting clutter. The procedure deals also with arrays containing NaN values. The map is the number of threshold exceeds within the window for this gate.

Parameters

array (xarray.DataArray) – The reflectivity array to operate on.

create_cluttermap(array=None, addname='')

Creates clutter map based on the field and threshold.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended cluttermaps overwrite the results.

Returns

cmap – The cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

ClutterMap

pylawr.transform.filter.SNR

class pylawr.transform.filter.SNR(threshold=0)

Bases: ClutterFilter

Calculates the Signal-to-Noise ratio.

threshold

The threshold of the reflectivities to filter (linear reflectivity)

Type

float

_abc_impl = <_abc._abc_data object>

calc_map(array, noise_remover)

Computes Signal-to-Noise ratio. The array should not corrected for beam expansion.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• noise_remover (NoiseRemover object) – Actual NoiseRemover which operates on RadarField

create_cluttermap(array=None, noise_remover=None)

Creates clutter map based on the field and threshold.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• noise_remover (NoiseRemover) – Actual NoiseRemover which operates on RadarField

Returns

cmap – The Cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

ClutterMap

pylawr.transform.filter.SpeckleFilter

class pylawr.transform.filter.SpeckleFilter(threshold=2, window_size=(3, 3))

Bases: ClutterFilter

The speckle filter checks the number of rain pixels in a neighborhood. If this number of rain pixels is lower than a given threshold, this pixel is declared as clutter. The boundary is zero-padded.

threshold

The threshold of the rain pixels in the neighborhood (default=2)

Type

int, optional

window_size

This window size is checked (default=(3, 3))

Type

tuple(int, int)

_abc_impl = <_abc._abc_data object>

calc_map(array)

This method estimates the number of rain pixels in the neighborhood and if the center pixel is also a rain pixel.

Parameters

array (xarray.DataArray) – This data array is used to determine the number of rain pixels and is used to estimate the raw cluttermap without threshold.

create_cluttermap(array=None, addname='')

Creates clutter map based on field, set neighborhood size and set threshold. If the number of rain pixels in the neighborhood is lower than given threshold, the center pixel is declared as clutter.

Parameters

• array (xarray.DataArray or None, optional) – This data array is used to determine the number of rain pixels and is used to estimate the cluttermap. If this is None (default), the stored clutter map is used.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended ClutterMaps overwrite the results.

Returns

cmap – The Cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

pylawr.transform.filter.clutter.ClutterMap

property window_product

pylawr.transform.filter.SPKFilter

class pylawr.transform.filter.SPKFilter(threshold=3.0, window_size=11, window_criterion=0.5, spike_width=1)

Bases: ClutterFilter

Calculates gradients of consecutive radar beams for clutter detection. If the difference between consecutive radar beams exceed a threshold for percentage (window_criterion) of consecutive range gates within a window, the gate is identified as clutter. This filter is similar to the RINGFilter, only the axis is changed.

With the default parameters a spike if six consecutive range gates is identified as clutter.

threshold

The threshold of the reflectivities to filter (dBZ)

Type

float

_wz

The number of consecutive range gates in which the threshold has to be exceeded

Type

int

_wc

The percentage of consecutive range gates the criteria have to be fulfilled

Type

float

_sw

The assumed spike width to detect.

Type

int

_abc_impl = <_abc._abc_data object>

calc_map(array)

Computes spike field for detecting clutter. The procedure deals also with arrays containing NaN values. The map is the number of threshold exceeds within the window for this gate.

Parameters

array (xarray.DataArray) – The reflectivity array to operate on.

create_cluttermap(array=None, addname='')

Creates clutter map based on the field and threshold.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended cluttermaps overwrite the results.

Returns

cmap – The cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

ClutterMap

pylawr.transform.filter.SPINFilter

class pylawr.transform.filter.SPINFilter(threshold=3, window_size=11, window_criterion=0.1)

Bases: ClutterFilter

Calculate SPIN change of the reflectivity for clutter detection according to Hubbert et al. (2009).

threshold

The threshold of the reflectivities to filter (dBZ)

Type

float

_wz

The number of consecutive range gates in which the threshold has to be exceeded

Type

int

_wc

The percentage of consecutive range gates the criteria have to be fulfilled

Type

float

_abc_impl = <_abc._abc_data object>

calc_map(array)

Computes SPIN field for detecting clutter. The SPIN feature field is a measure of how often the reflectivity gradient changes sign along the distance. The procedure deals also with arrays containing NaN values.

Parameters

array (xarray.DataArray) – The reflectivity array to operate on.

create_cluttermap(array=None, addname='')

Creates clutter map based on the field and window criteria. If both criteria of the SPIN filter are fulfilled in more than self._wc percent of the consecutive range gates, the centre range gate is flagged as clutter.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended cluttermaps overwrite the results.

Returns

cmap – The cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

ClutterMap

pylawr.transform.filter.TDBZFilter

class pylawr.transform.filter.TDBZFilter(threshold=3.0, window_size=5)

Bases: ClutterFilter

Calculate texture of reflectivity (TDBZ) for clutter detection according to Hubbert et al. (2009).

threshold

The threshold of the reflectivities to filter (dBZ)

Type

float

_wz

The number of consecutive range gates in which the threshold has to be exceeded

Type

int

_abc_impl = <_abc._abc_data object>

calc_map(array)

Computes TDBZ field for detecting clutter. The TDBZ field is computed as the average of the squared logarithmic reflectivity difference between adjacent range gates. The procedure deals also with arrays containing NaN values.

Parameters

array (xarray.DataArray) – The reflectivity array to operate on.

create_cluttermap(array=None, addname='')

Creates clutter map based on the field and threshold.

Parameters

• array (xarray.DataArray) – The reflectivity array to operate on.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended cluttermaps overwrite the results.

Returns

cmap – The cluttermap of _map to transform the reflectivity, contains probabilities of clutter (between one and zero).

Return type

ClutterMap

pylawr.transform.filter.TemporalFilter

class pylawr.transform.filter.TemporalFilter(store_n_images=3, threshold=3)

Bases: MemoryMixin, ClutterFilter

The temporal clutter filter compares current radar image with n-saved radar images. If the number of rain pixels within the last n-images is below a given threshold, then the pixel is identified as clutter. In this implementation fit and calc_map are split, because they are applied at different times: fit is normally applied before any clutter is removed and calc map is only used to identify current rain pixels and can be applied at any time, also after removing some clutter.

Parameters

• store_n_images (int, optional) – This number of images (default=3) is stored within this clutter filter.

• threshold (int, optional) – This threshold (default=3) is used to decide if a rain pixel is clutter or not. If the number of rain pixels within the stored images is below this threshold, the rain pixel is identified as clutter.

Notes

fit() should be normally applied before any other clutter filter was applied, while calc_map() can be used anywhere.

_abc_impl = <_abc._abc_data object>

calc_map(array)

This converts given array to rain pixels and sets this rain array as current map. This method set current map independent from fit().

Parameters

array (xarray.DataArray) – This array is converted into rain pixels.

create_cluttermap(array=None, addname='')

This method creates the cluttermap based on historic rain maps, current rain map and set threshold.

Parameters

• array (xarray.DataArray or None, optional) – If this array is given, then this will be converted into rain pixels and set as current map.

• addname (str) – Supplement to the name of the created ClutterMap (should be unique), otherwise appended ClutterMaps overwrite the results.

Returns

clutter_map – This is the cluttermap of this filter, dependent on stored images.

Return type

pylawr.transform.filter.cluttermap.ClutterMap

Notes

If the number of stored images is lower than the set attribute store_n_images, an empty clutter map will be returned.

fit(array)

This fit method sets the history maps, which are used to determine clutter. Also the maps are constrained to the last n-maps. This method is independent from calc_map().

Parameters

array (xarray.DataArray) – This array is concatenated to the historic maps.

property fitted

If the number of stored images equals the set number of stored images.

Returns

fitted – If the number of stored images equals the set number of stored images.

Return type

bool

to_xarray()

Serialize this filter’s parameters to an xarray.Dataset

Returns

the filter’s parameters as dataset

Return type

xarray.Dataset