Easily Generate Timeseries Ready InSAR Data With ASF Vertex
One of InSAR’s biggest hurdles is addressed. In a few clicks ASF Vertex can provide timeseries ready InSAR data using a free cloud-based service
InSAR has a capacity building problem. It’s a complicated process, it has high compute requirements, and high data storage requirements. This has been a hurdle for people that want to learn InSAR processing. How can you teach this type of methodology if students don’t have access to machines that are at least a thousand dollars and a good internet to download hundreds of gigabytes of data?
Recently, this barrier of entry has been addressed by Alaska Satellite Facility (ASF) because now they provide a new free cloud-based InSAR processing through their Vertex service. All you have to do is download the processed images and ingest them into your algorithm of choice. Currently only the small baseline subset (SBAS) algorithm is accepted which uses a network of images to analyze timeseries data.
In this post I’ll be focusing on exploring this service with MintPy which specializes in the SBAS timeseries algorithm. Currently there is no support for StaMPS which is another popular InSAR timeseries algorithm that uses the persistent scatterer method.
Getting the data
Ensure you have an account then you can do the following steps:
- Search for the data to get a scene ID from your AOI
2. Search for the SLC scene you want to use as reference then click “SBAS” in the scene detail panel. This will filter results according to your flight direction but it will also use images that don’t perfectly overlap (which is fine).
3. After filtering for your desired parameters such as date range, spatial baseline, and temporal baseline add all results to your “On-Demand Queue” using the InSAR GAMMA option.
4. In your on demand queue set up the parameters you want and set a project name so you can easily search for these files when you download them in the “Submitted Products” area. Make sure to include a DEM, Inc. Angle Map, and the wrapped phase. Those data are required for timeseries analysis with MintPy.
5. Once the processing is finished you can download them all at the same time using a Python script which is recommended. Otherwise you will be downloading them individually in the web browser.
Search for your processed data using the search bar above. Change search type to “On Demand Products” then type in your project name to filter the products:
ASF can process hundreds of GB of data in an hour or more depending on their servers. They process everything in parallel so even if you have hundreds of scenes it will all be done very quickly.
This data is ready for ingestion into MintPy which will do the final processing. Below are some of the outputs from the sample data I processed in London and Dubai just to see how it would perform. In these samples I processed 2 years worth of data in only a few hours.