https://github.com/orthoseg/orthoseg

OrthoSeg makes it easy to train neural networks to segment orthophotos.
https://github.com/orthoseg/orthoseg

aerial-imagery convolutional-neural-networks deep-learning drone geoai image-classification image-segmentation keras-tensorflow machine-learning neural-network python remote-sensing satellite-imagery segment-orthophotos segmentation tensorflow wms

Last synced: about 2 months ago
JSON representation

OrthoSeg makes it easy to train neural networks to segment orthophotos.

• Host: GitHub
• URL: https://github.com/orthoseg/orthoseg
• Owner: orthoseg
• License: gpl-3.0
• Created: 2018-09-05T11:25:22.000Z (about 7 years ago)
• Default Branch: main
• Last Pushed: 2025-08-18T06:52:20.000Z (2 months ago)
• Last Synced: 2025-08-18T08:39:08.832Z (2 months ago)
• Topics: aerial-imagery, convolutional-neural-networks, deep-learning, drone, geoai, image-classification, image-segmentation, keras-tensorflow, machine-learning, neural-network, python, remote-sensing, satellite-imagery, segment-orthophotos, segmentation, tensorflow, wms
• Language: Python
• Homepage: https://orthoseg.readthedocs.io/
• Size: 4.52 MB
• Stars: 35
• Watchers: 4
• Forks: 5
• Open Issues: 19
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.md
  • License: LICENSE.MD
  • Citation: CITATION.cff

Awesome Lists containing this project

README

          
# Orthophoto segmentation

[![Actions Status](https://github.com/orthoseg/orthoseg/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/orthoseg/orthoseg/actions?query=workflow%3ATests)

[![Coverage Status](https://codecov.io/gh/orthoseg/orthoseg/branch/main/graph/badge.svg)](https://codecov.io/gh/orthoseg/orthoseg)

[![PyPI version](https://img.shields.io/pypi/v/orthoseg.svg)](https://pypi.org/project/orthoseg)

[![DOI](https://zenodo.org/badge/147507046.svg)](https://zenodo.org/doi/10.5281/zenodo.10340584)

A python package that makes it (relatively) easy to segment orthophotos. Any type of

georeferenced images should work, e.g. satellite, aerial or drone images, (historical)

maps, hillshades,...

No programming is needed, everything is managed via configuration files.

The typical steps:

1. create a training dataset for a topic of your choice, e.g. in QGIS

2. train a neural network to segment orthophotos

3. run the segmentation on a larger area + vectorize the result

4. apply some basic postprocessing like dissolve, simplify,...

Only open source software is needed, eg. QGIS and tensorflow.

Installation and usage instructions can be found in the [orthoseg docs](https://orthoseg.readthedocs.io)

This is an example of how the output of a tree detection on aerial images can look:

![Result of a tree detection on aerial images](docs/_static/images/trees.jpg)