vhrharmonize: VHR Satellite Imagery Preprocessing Library

Important

This library is under active development and may change.

---

Overview

vhrharmonize is an open-source Python library and CLI suite for preprocessing very high resolution (VHR) satellite imagery into analysis-ready products. Current supported end-to-end workflows: WorldView-3 B1 imagery. Additional providers and sensors (for example, Planet) will be added.

---

Features

• Atmospheric correction workflows (Py6S default, FLAASH optional backend)
• RPC orthorectification with Orthority
• Pansharpening with Orthority
• Optional cloud masking with OmniCloudMask
• Pairwise alignment (coregix)
• Relative Radiometric Normalization with spectralmatch
• WorldView scene discovery, IMD parsing, and standardized metadata mapping
• CLI and library-first interfaces

---

Installation

See docs/getting-started/installation.md for detailed installation instructions or simply install like this:

conda create -n vhrharmonize -c conda-forge py6s sixs gdal python=3.11
conda activate vhrharmonize
pip install vhrharmonize[defaults]

Getting Started

For an overview of using the library see docs/getting-started/quickstart.md. The CLI can be usd by passing in arguments from a yaml file like this one docs/configs/worldview.example.yml and running:

vhr-worldview --config-yaml worldview.example.yml

Or pass in arguments directly from the command line:

vhr-worldview \
  --input-file-glob "/data/worldview/**/*.TIF" \
  --output-dir ../../processed \
  --run-alignment \
  --alignment-fixed-image /data/reference.tif

For detailed arguments use:

vhr-worldview --help
vhr-fetch-modis-water-vapor --help
vhr-flaash --help
vhr-cloudmask-raster --help
vhr-pansharpen-orthos --help
vhr-align-image-pair --help
vhr-orthorectification --help
vhr-radiometric-normalization --help
vhr-py6s --help

Contributing

We welcome all contributions! We appreciate any feedback, suggestions, or pull requests to improve this project. See docs/getting-started/contributing.md.

---

License

This project is licensed under the MIT License. See LICENSE for details.