We’re in St. Louis for FOSS4GNA


3 min read

ff003a783dfd  0 htTBGBKZuOWjIgv

At Development Seed, open source technology is central to how we think about and build our tools. We’re in St. Louis this week participating in one of the largest gatherings of open source contributors, adopters, and users at FOSS4GNA. We will share Label Maker, a python library that helps to extract insights from satellite imagery. We will present Urchn, a tool for mappers and planners to assess where changes are happening in urban areas, and where their map is out-of-date. Finally, we will discuss the current state of STAC (SpatioTemporal Asset Catalog), a common spec for metadata, and our work supporting this standard in our sat-utils libraries.

Tonight we’re sponsoring a Machine Learning Happy Hour in partnership with Climate Corporation, Azavea, and Astraea, which will highlight open source projects for machine learning and computer vision, and how you can apply them to your own work. I’ll give a quick lightning talk on Label Maker. Hope to see you tonight.

Over the course of the conference, we’ll be presenting and demoing some of our tools.

  • Matt will present Urchn, an urban change detection tool that shows you how your city is changing and helps you keep the map up-to-date. It uses machine learning methods to detect and extract changes in buildings, roads, green areas and other urban features, and uses these indicators to examine the development of regions over time. Urban Change Monitoring and OSM is today, Tuesday, May 15, 4:30pm — 5:05pm.
  • Matt will also discuss STAC, the SpatioTemporal Asset Catalog and the common metadata standards and API specifications to support the reuse of API and client software. The emerging STAC spec and sat-utils is Wednesday, May 16,2:45pm — 3:20pm.
  • I will talk about Label Maker which creates machine-learning-ready training data for most popular ML frameworks, including Keras, Tensor Flow, and MXNet. It pulls data from OpenStreetMap and combines that with imagery sources like Mapbox or Digital Globe to create a single file for use in training machine learning algorithms. Make your own Satellite Imagery Machine Learning Data is today, Tuesday, May 15, 2:00pm — 2:35pm.

Our colleagues at NASA and Element 84 are also presenting, and will share Cumulus, which leverages cloud computing for data processing, storage, and retrieval of their Earth observation data. How NASA is Building a Petabyte Scale Geospatial Archive in the Cloud is today, Tuesday, May 15, 12:00pm — 12:35pm.

Follow us throughout the week on Twitter for updates from FOSS4GNA. If you’re attending, ping Matt at @GeoSkeptic, Aimee at @_aimeeb, or me @drewbo19 to talk about machine learning, STAC, Cumulus, or other open source tools. And stop by our ML After Party tonight for a drink 🍺.

What we're doing.