# WangLibo1995/GeoSeg

UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery, ISPRS. Also, including other vision transformers and CNNs for satellite, aerial image and UAV image segmentation.

Repository: https://github.com/WangLibo1995/GeoSeg
Canonical: https://ross.abutalabs.com/products/geoseg
Language: Python
License: GPL-3.0
License Family: copyleft
Topics: pytorch-lightning, remote-sensing-image, vision-transformer, deep-learning, pytorch, timm, cnn, segmentation, semantic-segmentation
Last push: 2024-08-19T12:44:12+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1779, "days_push": 744, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1096, forks 153 (observed 2026-08-28T04:03:34.432598+00:00)

## What it is
GeoSeg is an open-source PyTorch-based semantic segmentation toolbox focused on Vision Transformers for remote sensing imagery, featuring the UNetFormer model. It supports training and inference on satellite, aerial, and UAV datasets like ISPRS Potsdam/Vaihingen, LoveDA, and UAVid.

## Use cases
- segment satellite images into land cover classes
- train UNetFormer on ISPRS Potsdam dataset
- semantic segmentation of UAV aerial imagery
- run inference on huge remote sensing images
- benchmark vision transformer segmentation models
- segment urban scenes from aerial photos

## When to choose
- you need state-of-the-art semantic segmentation for remote sensing or UAV imagery
- you want pretrained vision transformer backbones for aerial image segmentation
- you need multi-GPU training with PyTorch Lightning for segmentation benchmarks

## When to avoid
- you need general-purpose natural image segmentation rather than remote sensing
- you require a non-GPL license for commercial products
- you want a plug-and-play tool without writing training configs

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing
- platform: python
- tags: semantic-segmentation, remote-sensing, vision-transformer, pytorch-lightning, unetformer, uav-imagery, satellite-imagery, gpu

## Member repositories
- WangLibo1995/GeoSeg (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:34.432598+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:46:39.187665+00:00, confidence not recorded.
  - readme: https://github.com/WangLibo1995/GeoSeg (fetched 2026-08-28T04:03:34.432598+00:00, sha 2300bc6a7946)
- Data as of 2026-08-30T08:39:29.467469+00:00.