Mask¶

Specified vector or raster is interpreted as a boolean mask.

In case of a vector, all pixels covered by features are interpreted as True, all other pixels as False.

In case of a raster, all pixels that are equal to the no data value (default is 0) are interpreted as False, all other pixels as True.Multiband rasters are first evaluated band wise. The final mask for a given pixel is True, if all band wise masks for that pixel are True.

Invert Mask¶

Whether or not to invert the selected mask.

HTML Report¶

Specify output path for HTML report file (.html).

Specify output path for regression raster.

Output Fraction¶

Specify output path for fraction raster.

Open Test Maps¶

Opens testdata into current QGIS project (LandCov_BerlinUrbanGradient.shp, HighResolution_BerlinUrbanGradient.bsq, EnMAP_BerlinUrbanGradient.bsq, SpecLib_BerlinUrbanGradient.sli).

EnMAP (30m; 177 bands)¶

File name: EnMAP_BerlinUrbanGradient.bsq

Simulated EnMAP data (based on 3.6m HyMap imagery) acquired in August 2009 over south eastern part of Berlin covering an area of 4.32 km^2 (2.4 x 1.8 km). It has a spectral resolution of 177 bands and a spatial resolution of 30m.

HyMap (3.6m; Blue, Green, Red, NIR bands)¶

File name: HighResolution_BerlinUrbanGradient.bsq

HyMap image acquired in August 2009 over south eastern part of Berlin covering an area of 4.32 km^2 (2.4 x 1.8 km). This dataset was reduced to 4 bands (0.483, 0.558, 0.646 and 0.804 micrometers). The spatial resolution is 3.6m.

LandCover Layer¶

File name: LandCov_BerlinUrbanGradient.shp

Polygon shapefile containing land cover information on two classification levels. Derived from very high resolution aerial imagery and cadastral datasets.

Level 1 classes: Impervious; Other; Vegetation; Soil

Level 2 classes: Roof; Low vegetation; Other; Pavement; Tree; Soil

Library as Raster¶

File name: SpecLib_BerlinUrbanGradient.sli

Spectral library with 75 spectra (material level, level 2 and level 3 class information)

Open Test Library¶

Unique Values from Raster Band¶

This algorithm returns unique values from a raster band as a list. The output will be shown in the log window and can the copied from there accordingly.

Raster¶

Specify input raster.

Band¶

Specify input raster band.

View Raster Metadata¶

Prints all Raster metadata to log.

Raster¶

Specify input raster.

Unique Values from Vector Attribute¶

This algorithm returns unique values from vector attributes as a list, which is also usable as Class Definition in other algorithms. The output will be shown in the log window and can the copied from there accordingly.

Vector¶

Specify input vector.

Field¶

Specify field of vector layer for which unique values should be derived.

Classification¶

Fit GaussianProcessClassifier¶

Fits Gaussian Process Classifier. See Gaussian Processes for further information.

Raster¶

Specify input raster.

Classification¶

Specify input raster.

Mask¶

Specified vector or raster is interpreted as a boolean mask.

In case of a vector, all pixels covered by features are interpreted as True, all other pixels as False.

Classification from Fraction¶

Creates classification from class fraction. Winner class is equal to the class with maximum class fraction.

ClassFraction¶

Specify input raster.

Minimal overall coverage¶

Mask out all pixels that have an overall coverage less than the specified value. This controls how edges between labeled and no data regions are treated.

Minimal dominant coverage¶

Mask out all pixels that have a coverage of the predominant class less than the specified value. This controls pixel purity.

Output Classification¶

Specify output path for classification raster.

Classification from Vector¶

Creates a classification from a vector field with class ids.

Pixel Grid¶

Specify input raster.

Vector¶

Specify input vector.

Class id attribute¶

Vector field specifying the class ids.

Class Definition¶

Enter a class definition, e.g.:

ClassDefinition(names=[‘Urban’, ‘Forest’, ‘Water’], colors=[‘red’, ‘#00FF00’, (0, 0, 255)])

For supported named colors see the W3C recognized color keyword names.

Minimal overall coverage¶

Mask out all pixels that have an overall coverage less than the specified value. This controls how edges between labeled and no data regions are treated.

Minimal dominant coverage¶

Mask out all pixels that have a coverage of the predominant class less than the specified value. This controls pixel purity.

Oversampling factor¶

Defines the degree of detail by which the class information given by the vector is rasterized. An oversampling factor of 1 (default) simply rasterizes the vector on the target pixel grid.An oversampling factor of 2 will rasterize the vector on a target pixel grid with resolution twice as fine.An oversampling factor of 3 will rasterize the vector on a target pixel grid with resolution three times as fine, … and so on.

Mind that larger values are always better (more accurate), but depending on the inputs, this process can be quite computationally intensive, when a higher factor than 1 is used.

Output Classification¶

Specify output path for classification raster.

Fraction from Classification¶

Derive (binarized) class fractions from a classification.

Classification¶

Specify input raster.

Output Fraction¶

Specify output path for fraction raster.

Fraction from Vector¶

Derives class fraction raster from a vector file with sufficient class information.

Pixel Grid¶

Specify input raster.

Vector¶

Specify input vector.

Class id attribute¶

Vector field specifying the class ids.

Class Definition¶

Enter a class definition, e.g.:

ClassDefinition(names=[‘Urban’, ‘Forest’, ‘Water’], colors=[‘red’, ‘#00FF00’, (0, 0, 255)])

For supported named colors see the W3C recognized color keyword names.

Minimal overall coverage¶

Mask out all pixels that have an overall coverage less than the specified value. This controls how edges between labeled and no data regions are treated.

Minimal dominant coverage¶

Mask out all pixels that have a coverage of the predominant class less than the specified value. This controls pixel purity.

Oversampling factor¶

Defines the degree of detail by which the class information given by the vector is rasterized. An oversampling factor of 1 (default) simply rasterizes the vector on the target pixel grid.An oversampling factor of 2 will rasterize the vector on a target pixel grid with resolution twice as fine.An oversampling factor of 3 will rasterize the vector on a target pixel grid with resolution three times as fine, … and so on.

Mind that larger values are always better (more accurate), but depending on the inputs, this process can be quite computationally intensive, when a higher factor than 1 is used.

Output Fraction¶

Specify output path for fraction raster.

Raster from Vector¶

Converts vector to raster (using gdal rasterize).

Pixel Grid¶

Specify input raster.

Vector¶

Specify input vector.

Init Value¶

Pre-initialization value for the output raster before burning. Note that this value is not marked as the nodata value in the output raster.

Burn Value¶

Fixed value to burn into each pixel, which is covered by a feature (point, line or polygon).

Burn Attribute¶

Specify numeric vector field to use as burn values.

All touched¶

Enables the ALL_TOUCHED rasterization option so that all pixels touched by lines or polygons will be updated, not just those on the line render path, or whose center point is within the polygon.

Filter SQL¶

Create SQL based feature selection, so that only selected features will be used for burning.

Example: Level_2 = ‘Roof’ will only burn geometries where the Level_2 attribute value is equal to ‘Roof’, others will be ignored. This allows you to subset the vector dataset on-the-fly.

Data Type¶

Specify output datatype.

No Data Value¶

Specify output no data value.

Output Raster¶

Mask¶

Specified vector or raster is interpreted as a boolean mask.

In case of a vector, all pixels covered by features are interpreted as True, all other pixels as False.

Specify output path for raster.

Output Fraction¶

Specify output path for fraction raster.

Masking¶

Build Mask from Raster¶

Builds a mask from a raster based on user defined values and value ranges.

Raster¶

Specify input raster.

Foreground values¶

List of values and ranges that are mapped to True, e.g. [1, 2, 5, range(5, 10)].

Background values¶

List of values and ranges that are mapped to False, e.g. [-9999, range(-10, 0)].

Output Mask¶

Specify output path for mask raster.

Apply Mask to Raster¶

Pixels that are masked out are set to the raster no data value.

Raster¶

Specify input raster.

Mask¶

Specified vector or raster is interpreted as a boolean mask.

In case of a vector, all pixels covered by features are interpreted as True, all other pixels as False.

In case of a raster, all pixels that are equal to the no data value (default is 0) are interpreted as False, all other pixels as True.Multiband rasters are first evaluated band wise. The final mask for a given pixel is True, if all band wise masks for that pixel are True.

Invert Mask¶

Whether or not to invert the selected mask.

Masked Raster¶

Specify output path for raster.

Post-Processing¶

Fraction as RGB Raster¶

Creates a RGB representation from given class fractions. The RGB color of a specific pixel is the weighted mean value of the original class colors, where the weights are given by the corresponding class propability.

ClassFraction¶

Specify input raster.

Output Raster¶

Specify output path for raster.

Random¶

Random Points from Classification¶

Randomly samples a user defined amount of points/pixels from a classification raster and returns them as a vector dataset.

Classification¶

Specify input raster.

Number of Points per Class¶

Has to be a number or a list of numbers. When a single integer number is given (e.g. 100), equalised random sample will be taken, i.e. in this case 100 samples per class. For taking a disproportional random sample, where the amount of samples should differ between classes, provide a list of numbers. This list has to have as many arguments as classes in the classification and has to be ordered according to the classes (e.g. ‘[100, 70, 90]’ in the case of three classes, 100 samples will be taken from the first class, 70 from the second, etc.). For a proportional stratified random sampling provide a float value between 0 and 1 (e.g. 0.3 for randomly drawing 30% of pixels in each class).

Output Vector¶

Specify output path for the vector.

Random Points from Mask¶

Randomly draws defined number of points from Mask and returns them as vector dataset.

Mask¶

Specified vector or raster is interpreted as a boolean mask.

In case of a vector, all pixels covered by features are interpreted as True, all other pixels as False.

In case of a raster, all pixels that are equal to the no data value (default is 0) are interpreted as False, all other pixels as True.Multiband rasters are first evaluated band wise. The final mask for a given pixel is True, if all band wise masks for that pixel are True.

Invert Mask¶

Whether or not to invert the selected mask.

Number of Points¶

Number of points to sample from mask.

Output Vector¶

Specify output path for the vector.

Select path to a regressor file (.pkl).

Output Regression¶

Specify output path for regression raster.

Resampling¶

Specify output path for raster.