Build STAC for your geospatial data
Introduction
The Azure Space team has just released a new reference architecture, “Organize spaceborne geospatial data with SpatioTemporal Asset Catalog”. This guide is the first of its kind, providing step-by-step guidance on how to acquire geospatial data, generate associated metadata, create STAC catalogs, and then query against these catalogs via STAC APIs. Along with the reference architecture guide, the team has also published the sample code on GitHub to implement STAC based on public sample data from the National Agriculture Imagery Program (NAIP). The sample code can be customized and adapted for your own data sets. It is built on the Azure platform to leverage Azure’s rich product ecosystem with built-in security, monitoring, and scalability capabilities at low cost. Since the STAC implementation uses Azure's core platform, it can be complemented with other Azure services to optimize scale, throughput and time to insights.
Background
In the past, every geospatial data provider had its own standard and protocol/APIs to allow developers and users access to its geospatial data. The SpatioTemporal Asset Catalog (STAC) family of specifications aims to standardize the way geospatial asset metadata is structured and queried. The goal of the common standard is to eliminate the need for many APIs across multiple satellite providers. With its first version released in 2017 by a community of geospatial developers from 14 organizations including Radiant.Earth, Planet, and DigitalGlobe, STAC has become an industry standard to organize and describe geospatial information, so it can easily be worked with, indexed, and discovered. Today, STAC has been deployed in many production systems such as Microsoft Planetary Computer, Planet Orders API, Google Earth Engine, and others.
Why it matters to me
If you are a geospatial data provider, developer, analyst, or user who has one or many data sets that originate from various data sources such as a new satellite, a commercial data provider, or your own private drone captured data, you will benefit from building STAC catalogs in these three areas:
- Application and tooling parity. All geospatial data sets are organized in the same way so that your own or your customers’ applications don’t need to change the way they interact with previous data sets. All of your existing GIS and open-source software (OSS) tooling still function the same way, so the newly added data sets will not break these tools.
- Support of a variety of data types. Your application may need to access various types of geospatial data such as optical imagery, synthetic aperture radar (SAR), point clouds, lidar, digital elevation models (DEM), vector, and so forth. STAC, being the common query/access interface for all these different types of data, makes developing such applications with data aggregation and fusion needs much easier and more efficient.
- Performance improvement. Caching a large amount of geospatial data in STAC catalog(s) close to your application provides improved performance and repeatable workflows. When your application needs to process thousands of GeoTiff files, for example, as part of a repeatable workflow such as object detection analysis with various configurations, instead of downloading these files repeatedly from a remote data source, creating STAC catalogs for these files locally would help save network bandwidth and boost performance of your application as well.
Next steps
Learn more details to organize spaceborne geospatial data with SpatioTemporal Asset Catalog and deploy the sample solution in your Azure account today. Creating STAC catalogs for your geospatial data unlocks insights from your valuable data estate. To learn about making connection to your satellite and downlink live data, please refer to “Tutorial: Collect and process Aqua satellite data using Azure Orbital Ground Station (AOGS)”. Also check out “Spaceborne data analysis with Azure Synapse Analytics” for guidance on an end-to-end implementation that involves extracting, loading, transforming, and analyzing spaceborne data by using geospatial libraries and AI models with Azure Synapse Analytics.
References
