Skip to content

Latest commit

 

History

History
173 lines (121 loc) · 5.62 KB

README.md

File metadata and controls

173 lines (121 loc) · 5.62 KB
Raw

UKIS UKIS_sat2h5

ukis_sat2h5 can be used to convert many (similar sized) GeoTiff images into a HDF-5 array while keeping track of spatial extent and projections. HDF-5 arrays are easier to use in deep learning applications. All images will be padded to fit the same dimensions. ukis_sat2h5 also includes tools for tiling images (with optional overlap) and converting images back to GeoTiff files.

Workflow overview

Imagine a folder structure like this:

images
├── image1
│   ├── image.tif
│   └── label.tif
├── image2
│   ├── image.tif
│   └── label.tif
└── image3
    ├── image.tif
    └── label.tif

ukis_sat2h5 will convert this into a HDF-5 file with the following structure:

images.hdf5
  /affine    : Dataset {<nimages>, 6}
  /epsg      : Dataset {<nimages>}
  /img       : Dataset {<nimages>, <nbands>, <nrows>, <ncols>}
  /img_means : Dataset {<nbands>}
  /img_stds  : Dataset {<nbands>}
  /lbl       : Dataset {<nimages>, <nrows>, <ncols>}
  /path      : Dataset {<nimages>}

Examples

$ python -m ukis_sat2h5 img_to_h5 -r ~/tmp/sen2_ref_images -d ~/tmp/sen2_ref_images.h5 -f "*.vrt" -l "*label*" -b 2 3 4 8
$ python -m ukis_sat2h5 tile_h5 -s ~/tmp/sen2_ref_images.h5 -d ~/tmp/sen2_ref_images_tiled.h5 -t 256 -o 128 -a 1024
$ python -m ukis_sat2h5 h5_to_image -s ~/tmp/sen2_ref_images_tiled.h5 -d ~/tmp/tiles_to_images -i 1 42 128

Installation

This project uses a conda environment. For installing dependencies use:

conda env create

Dependencies

For the latest list of dependencies check the environment.yml.

Usage

CLI

This package provides standalone CLI functionality

usage: ukis_sat2h5 [-h] {img_to_h5,h5_to_img,tile_h5} ...

Convert geographic images to H5 file and back and tile them for Deep Learning applications

positional arguments:
  {img_to_h5,h5_to_img,tile_h5}
                        Tools
    img_to_h5           Convert geographic images to H5 arrays
    h5_to_img           Convert H5 arrays (created with this tool) to back to geographic images
    tile_h5             Tile H5 array and preserve geographic location

options:
  -h, --help            show this help message and exit

Python Module

The main functionality is encapsulated inside the follwing functions

import ukis_sat2h5

ukis_sat2h5.conversion.convert_img_to_h5(src_path, dst_file, file_glob, image_bands)
ukis_sat2h5.conversion.convert_h5_to_img(src_file, dst_folder, index)
ukis_sat2h5.tiling.tile_h5(h5_file, h5_file_tiled, tile_size, overlap, target_size)

Development

pre-commit

Some development guardrails are enforced via pre-commit. This is to ensure we follow similar code styles or it automatically cleans up jupyter notebooks.

To install pre-commit (not necessary if you installed the conda environment):

$ # conda/pip install pre-commit

To initialize all pre-commit hooks, run:

$ pre-commit install

To test whether pre-commit works:

$ pre-commit run --all-files

It will check all files tracked by git and apply the triggers set up in .pre-commit-config.yaml. That is, it will run triggers, possibly changing the contents of the file (e.g. black formatting). Once set up, pre-commit will run, as the name implies, prior to each git commit. In its current config, it will format code with black and isort, clean up jupyter notebook output cells, remove trailing whitespaces and will block large files to be committed. If it fails, one has to re-stage the affected files (git add or git stage), and re-commit.

pytest and coverage

We use pytest for testing the code and coverage.py to analyze code coverage:

$ # pip install pytest coverage
$ coverage erase;
$ coverage run --concurrency=multiprocessing -m pytest
$ coverage combine # requried as we use multiprocessing
$ coverage report
$ coverage html

Contributors

The UKIS team creates and adapts libraries which simplify the usage of satellite data. Our team includes (in alphabetical order):

  • Weigand, Matthias

German Aerospace Center (DLR)

Licenses

This software is licensed under the Apache 2.0 License.

Copyright (c) 2023 German Aerospace Center (DLR) * German Remote Sensing Data Center * Department: Geo-Risks and Civil Security

Changelog

See changelog.

Contributing

The UKIS team welcomes contributions from the community. For more detailed information, see our guide on contributing if you're interested in getting involved.

What is UKIS?

The DLR project Environmental and Crisis Information System (the German abbreviation is UKIS, standing for Umwelt- und Kriseninformationssysteme aims at harmonizing the development of information systems at the German Remote Sensing Data Center (DFD) and setting up a framework of modularized and generalized software components.

UKIS is intended to ease and standardize the process of setting up specific information systems and thus bridging the gap from EO product generation and information fusion to the delivery of products and information to end users.

Furthermore, the intention is to save and broaden know-how that was and is invested and earned in the development of information systems and components in several ongoing and future DFD projects.