DANSAR (Data Applications Notebooks with Synthetic Aperture Radar)

The Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) is a facility instrument suite built and maintained at JPL. The L-band instrument is a testbed for NISAR (NASA ISRO Synthetic Aperture Radar), a spaceborne instrument planned to launch in 2023. UAVSAR observations are operationally processed at JPL and can be downloaded by the public from the Alaska Satellite Facility. More information can be found here: https://uavsar.jpl.nasa.gov/

The last decade has experienced a significant increase in the demand and utilization of SAR imagery for various science applications. The Data Applications Notebooks with SAR (DANSAR) described here use concrete examples to expose the broader remote sensing community to SAR datasets. By analyzing real observations, users should be able to appreciate some of the main characteristics of SAR datasets, its potential uses, as well as limitations. DANSAR Notebooks were designed to overcome two bottlenecks experienced by beginner users. First, notebooks are run in a cloud environment, removing the need for downloading images, setting up files, or installing software. Second, notebooks are divided by discipline allowing users to get specific examples covering problems they are familiar with. 

The routines can be implemented in Google Colab: https://colab.research.google.com/notebooks/intro.ipynb?utm_source=scs-index Users can connect and view without the need for a google account or can run the notebooks locally using the repo’s environment files. The Google Colab public interface provides access to processing and memory necessary to store and analyze UAVSAR images as well as ancillary files. We use a python 3.8 kernel and include steps for installing python libraries via pip. The routines use open-source GIS libraries such as numpy, geopandas, and GDAL.

The public GitHub Repository “DANSAR” hosts the collection of Google Colab Notebooks. To view the Google Colab Notebooks for each SAR application, users may navigate to the Jupyter Notebook (*.ipynb) files in the main branch of the repository. Notebooks are named for their SAR application. At the top of the selected notebook, there will be an "Open in Colab" button such as this image here , which is a link replacing the web page URL "github.com" with the string "colab.research.google.com". Clicking this button at the top of the *.ipynb notebook will launch a new webpage with the Google Colab Notebook. A Google Account login is required to execute and upload data to the notebooks.

The main inputs for the notebooks will be UAVSAR imagery downloaded from the Alaska Satellite Facility and cropped to reduce file size. We will also consider ancillary datasets to help with interpretation. For example, maps of forest fire scars from the National Interagency Fire Center, basemaps showing administrative units and roads, and maps of known fault lines in California. These data files are accessible from this JPL website: https://uavsar.jpl.nasa.gov/cgi-bin/sar-notebooks.pl. They can either be imported directly to the Google Colab Notebooks or first downloaded to run locally. 

Routines are maintained and stored from the public GitHub repository DANSAR: https://github.com/anniepeacock/DANSAR. DANSAR is distributed with an Apache license. Google Colab can pull routines from any public GIT repo and will also be used to update the code and generate new version files. Making changes in the browser is preferable as it ensures correct setup of the python environment.

URS Clearance Number CL#21-5401 © 2021 California Institute of Technology. Government sponsorship acknowledged.