i.sentinel_2.autotraining

NAME

i.sentinel_2.autotraining - Automatically generates training data from input bands and indices a reference classification and treecover map. Creates classes water, low vegetation, forest, bare soil and built-up.

KEYWORDS

imagery, satellite, Sentinel, classification, extraction

SYNOPSIS

i.sentinel_2.autotraining

i.sentinel_2.autotraining --help

i.sentinel_2.autotraining ndvi=name ndwi=name ndbi=name [bsi=name] ref_classification_probav=name ref_treecover_fraction_probav=name [ref_classification_gong=name] ref_ghs_built=name [percentage_threshold=float] output_vector=name output_raster=name [str_column=string] [int_column=string] npoints=integer [--overwrite] [--help] [--verbose] [--quiet] [--ui]

Flags:

--overwrite: Allow output files to overwrite existing files
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

ndvi=name [required]: Input NDVI raster map
ndwi=name [required]: Input NDWI raster map
ndbi=name [required]: Input NDBI raster map
bsi=name: Input BSI raster map
ref_classification_probav=name [required]: Input reference probav classification map
ref_treecover_fraction_probav=name [required]: Input reference probav treecover fraction map
ref_classification_gong=name: Input reference gong et al. classification map
ref_ghs_built=name [required]: Input reference global human settlement builtup map (GHS-BUILT)
percentage_threshold=float: Minimum percentage of area potential training data of a class has to cover to be included in the classification; Options: 0.0-50.0; Default: 0.1
output_vector=name [required]: Output vector map with training data points. Class information will be stored in columns str_column and int_column
output_raster=name [required]: Output raster map with potential training areas
str_column=string: Name of the string column in output_vector to store class information; Default: lulc_class_str
int_column=string: Name of the integer column in output_vector to store class information; Default: lulc_class_int
npoints=integer [required]: Number of sampling points per class in the output vector map; Default: 10000

DESCRIPTION
EXAMPLE
SEE ALSO
AUTHOR

DESCRIPTION

i.sentinel_2.autotraining is a GRASS addon that automatically creates raster and vector training/validation data for further classification. It covers the classes water, low vegetation, forest, bare soil and built-up.

EXAMPLE

i.sentinel_2.autotraining ndvi=ndvi ndwi=ndwi ndbi=ndbi bsi=bsi \
  ref_classification=PROBAV_classification_Borneo \
  ref_treecover_fraction=PROBA_V_100m_treecoverfraction \
  percentage_threshold=0.1 npoints=1000 output_vector=test_autotraining_vector \
  output_raster=test_autotraining_raster

AUTHOR

Guido Riembauer, mundialis GmbH & Co. KG

SOURCE CODE

Available at: i.sentinel_2.autotraining source code (history)

Accessed: Monday Jul 27 12:22:42 2026