GRASS logo

NAME

m.neural_network - GRASS Reference Manual

Table of contents

image-alt

NAME

m.neural_network - GRASS GIS addons to prepare data, train and apply, and postprocess data for neural network.

KEYWORDS

raster,
vector

DESCRIPTION

The m.neural_network toolset consists of several modules.

Data preparation:
- m.neural_network.tindex: Creates a tile index of the possible tiles for
the data preparation either in apply or training mode
- m.neural_network.worker_nullcells:
Worker to analyse the number of null cells of imagery data for training mode
- m.neural_network.preparedata_part1:
Prepares and exports tiles for the label process
- m.neural_network.preparedata_part1.worker_export:
Worker for parallel processing for exporting for
m.neural_network.preparedata_part1
- m.neural_network.preparedata_part2:
Prepares imagery and labelled data for training and application of a neural network.
- m.neural_network.preparedata_part2.worker_label:
Worker to rasterize labelled data in parallel for
m.neural_network.preparedata_part2

Model development and usage:
- m.neural_network.train: training of a
semantic segmentation model with smp libraries
- m.neural_nework.test: calculation of
statistics for quality assessment
- m.neural_network.apply: application of a
trained model to new data

Postprocessing:
- m.neural_network.postprocessing.patch: patches the tiles (GeoTIFFs) which results from neural network inference.
- m.neural_network.postprocessing.vectorize vectorizes the classification raster output and clean results: remove small areas, if set straighten lines.
- m.neural_network.postprocessing.snapref: merges classification vector with reference data.

REQUIREMENTS

v.out.geojson from mundialis.
For required python libraries see requirements_all.txt.

This includes:
- GDAL/OGR and Python bindings
- pytorch
- segmentation_models.pytorch
- pytorch-lightning
- albumentations

AUTHORS

Anika Weinmann, mundialis GmbH & Co. KG
Guido Riembauer, mundialis GmbH & Co. KG
Markus Metz, mundialis GmbH & Co. KG
Victoria-Leandra Brunn, mundialis GmbH & Co. KG
Lina Krisztian, mundialis GmbH & Co. KG

SOURCE CODE

Available at: m.neural_network source code (history)

Accessed: Sunday Nov 15 11:04:30 2026


Main index | Miscellaneous index | Topics index | Keywords index | Graphical index | Full index

© 2003-2026 GRASS Development Team, GRASS 8.5.1dev Reference Manual