Spatial convolution Savitsky-Golay filter

2D Savitsky-Golay filter. See wikipedia for details.

The Savitzky-Golay filter is a linear filter that performs convolution using a sliding window over the input data. It aims to smooth the data while preserving important features, such as edges or peaks, by fitting a polynomial function within the window and using it to estimate the filtered value at the center point.

In the case of a Spatial convolution 2D Savitzky-Golay filter, the filter operates on a two-dimensional image. It applies the Savitzky-Golay technique separately along each dimension (rows and columns) of the image.

The main purpose of the 2D Savitzky-Golay filter is to reduce noise and smooth the image while preserving the sharpness and fine details. By fitting a polynomial function to the local neighborhood of each pixel, it provides a weighted average that is influenced by the neighboring pixels, effectively reducing the noise contribution.

Usage:

1. Start the algorithm from the Processing Toolbox panel.

2. Select the raster to process and modify the parameterization if necessary, then click run.

   

3. Processed image in comparison to the original.

   

Parameters

Raster layer [raster]

Raster layer to be filtered.

Kernel [string]

Python code. See sgolay2d from the SciPy cookbook for information on different parameters.

Default:

from astropy.convolution import Kernel2D
from enmapboxprocessing.algorithm.spatialconvolutionsavitskygolay2dalgorithm import sgolay2d
kernel = Kernel2D(array=sgolay2d(window_size=11, order=3, derivative=None))

Normalize kernel [boolean]

Whether to normalize the kernel to have a sum of one.

Default: False

Interpolate no data pixel [boolean]

Whether to interpolate no data pixel. Will result in renormalization of the kernel at each position ignoring pixels with no data values.

Default: True

Outputs

Output raster layer [rasterDestination]

Raster file destination.

Command-line usage

>qgis_process help enmapbox:SpatialConvolutionSavitskygolayFilter:

----------------
Arguments
----------------

raster: Raster layer
    Argument type:  raster
    Acceptable values:
            - Path to a raster layer
kernel: Kernel
    Default value:  from astropy.convolution import Kernel2D
from enmapboxprocessing.algorithm.spatialconvolutionsavitskygolay2dalgorithm import sgolay2d
kernel = Kernel2D(array=sgolay2d(window_size=11, order=3, derivative=None))
    Argument type:  string
    Acceptable values:
            - String value
            - field:FIELD_NAME to use a data defined value taken from the FIELD_NAME field
            - expression:SOME EXPRESSION to use a data defined value calculated using a custom QGIS expression
normalize: Normalize kernel
    Default value:  false
    Argument type:  boolean
    Acceptable values:
            - 1 for true/yes
            - 0 for false/no
            - field:FIELD_NAME to use a data defined value taken from the FIELD_NAME field
            - expression:SOME EXPRESSION to use a data defined value calculated using a custom QGIS expression
interpolate: Interpolate no data pixel
    Default value:  true
    Argument type:  boolean
    Acceptable values:
            - 1 for true/yes
            - 0 for false/no
            - field:FIELD_NAME to use a data defined value taken from the FIELD_NAME field
            - expression:SOME EXPRESSION to use a data defined value calculated using a custom QGIS expression
outputRaster: Output raster layer
    Argument type:  rasterDestination
    Acceptable values:
            - Path for new raster layer

----------------
Outputs
----------------

outputRaster: <outputRaster>
    Output raster layer