DELTA (Deep Earth Learning, Tools, and Analysis) is a framework for deep learning on satellite imagery, based on Tensorflow. DELTA classifies large satellite images with neural networks, automatically handling tiling large imagery.

DELTA is currently under active development by the NASA Ames Intelligent Robotics Group. Initially, it is mapping floods for disaster response, in collaboration with the U.S. Geological Survey, National Geospatial Intelligence Agency, National Center for Supercomputing Applications, and University of Alabama.

Installation

1. Install python3, GDAL, and the GDAL python bindings. For Ubuntu Linux, you can run scripts/setup.sh from the DELTA repository to install these dependencies.

2. Install Tensorflow following the instructions. For GPU support in DELTA (highly recommended) follow the directions in the GPU guide.

3. Checkout the delta repository and install with pip:

   git clone http://github.com/nasa/delta
   python3 -m pip install delta

DELTA is now installed and ready to use!

Alternate install using a Conda environment

1. Install Miniconda.

2. Checkout the delta repository and cd into the directory:

   git clone http://github.com/nasa/delta
   cd ./delta
   

3. Run the following commands to create a conda environment, install dependences, and install delta:

   conda create --name delta --yes python gdal
   conda activate delta
   pip install . # This command should be run in the ./delta directory we made in step 2

Installing ESA SNAP

In order to process Sentinel-1 images you will need to instal the ESA SNAP tool. You can download it here: https://step.esa.int/main/download/snap-download/

Once it is installed, you will need to add SNAP's bin folder to your path like this: export PATH=$PATH:/where/you/installed/snap/bin

Documentation

DELTA can be used either as a command line tool or as a python library. See the python documentation for the master branch here, or generate the documentation with scripts/docs.sh.

Example

As a simple example, consider training a neural network to map clouds with Landsat-8 images. The script scripts/example/l8_cloud.sh trains such a network using DELTA from the USGS SPARCS dataset, and shows how DELTA can be used. The steps involved in this, and other, classification processes are:

1. Collect training data. The SPARCS dataset contains Landsat-8 imagery with and without clouds.

2. Label training data. The SPARCS labels classify each pixel according to cloud, land, water and other classes.

3. Train the neural network. The script scripts/example/l8_cloud.sh invokes the command

   delta train --config l8_cloud.yaml l8_clouds.h5
   

   where scripts/example/l8_cloud.yaml is a configuration file specifying the labeled training data and training parameters (learn more about configuration files below). A neural network file l8_clouds.h5 is output.

4. Classify with the trained network. The script runs

   delta classify --config l8_cloud.yaml --image-dir ./validate --overlap 32 l8_clouds.h5
   

   to classify the images in the validate folder using the network l8_clouds.h5 learned previously. The overlap tiles to ignore border regions when possible to make a more aesthetically pleasing classified image. The command outputs a predicted image and confusion matrix.

The results could be improved--- with more training, more data, an improved network, or more--- but this example shows the basic usage of DETLA.

Configuration and Extensions

DELTA provides many options for customizing data inputs and training. All options are configured via YAML files. Some options can be overwritten with command line options (use delta --help to see which). See the delta.config README to learn about available configuration options.

DELTA can be extended to support custom neural network layers, image types, preprocessing operations, metrics, losses, and training callbacks. Learn about DELTA extensions in the delta.config.extensions documentation.

Data Management

DELTA integrates with MLFlow to track training. MLFlow options can be specified in the corresponding area of the configuration file. By default, training and validation metrics are logged, along with all configuration parameters. The most recent neural network is saved to a file when the training program is interrupted or completes.

View all the logged training information through mlflow by running::

  delta mlflow_ui

and navigating to the printed URL in a browser. This makes it easier to keep track when running experiments and adjusting parameters.

Data Preparation

By default DELTA operates on compressed input images which are unpacked to a temporary cache before they are processed. You can speed up processing by pre-unpacking your input data to a new folder using the tool scripts/fetch/unpack_inputs.py as in this example:

  python3 scripts/fetch/unpack_inputs.py --input-folder raw_images --output-folder unpacked_images \
  --image-type worldview --image-ext .zip

The images will be unpacked in the output folder, ready for training or classification. To train or classify with unpacked data, the image type specified in the configuration file remains the same but the extension should match the new extension in the unpacked folders (.tif for worldview, .vrt for Sentinel1).

Contributors

We welcome pull requests to contribute to DELTA. However, due to NASA legal restrictions, we must require that all contributors sign and submit a NASA Individual Contributor License Agreement. You can scan the document and submit via email. Thank you for your understanding.

Important notes for developers:

• Branching: Active development occurs on develop. Releases are pushed to master.

• Code Style: Code must pass our linter before merging. Run scripts/linter/install_linter.sh to install the linter as a git pre-commit hook.

• Unit Tests: Code must pass unit tests before merging. Run pytest in the tests directory to run the tests. Please add new unit tests as appropriate.

• Development Setup: You can install delta using pip's -e flag which installs in editable mode. Then you can run delta and it will use your latest changes made to the repo without reinstalling.

Licensing

DELTA is released under the Apache 2 license.

Copyright (c) 2020, United States Government, as represented by the Administrator of the National Aeronautics and Space Administration. All rights reserved.