Founded by IQT Labs’ CosmiQ Works with Maxar Technologies in 2016, SpaceNet began as an informal collaboration to accelerate open-source machine learning capabilities specifically for geospatial use cases. Over time, SpaceNet grew into a robust collaboration among co-founder and managing partner CosmiQ Works, co-founder Maxar Technologies, and six valuable partners. CosmiQ Works ran the organization for the past five years and managed seven unique data science challenges focused on foundational mapping (building footprints and road networks). Among other applications, these challenges inform a number of humanitarian and national security use cases where maps may be outdated or lacking. This blog is an abridged version of the full post appearing on The DownLinQ.
In March 2021, IQT Labs and CosmiQ Works will step back from its leadership of SpaceNet and focus resources on new initiatives. CosmiQ Works will fold into IQT Labs. With CosmiQ Works’ exit, leadership of SpaceNet will be transitioned to Maxar. SpaceNet resources (data, code and website) will continue to be openly available. Before we go, we want to celebrate highlights and accomplishments that SpaceNet produced so far:
- Demonstrated that open-source challenges provide value for the U.S. government. For example, the advancements in road extraction and optimized routing resulting from SpaceNet Challenges 3 and 5 connect immediately to numerous national security missions.
- Made massive quantities of high-quality imagery and labeled data available to academia, industry, and government with a permissive license, including: ~67,000 square km of high-resolution imagery, >11m labeled building footprints, and ~20,000 km of road labels (see spacenet.ai for further details).
- Hosted seven data science challenges with worldwide participation, each focused on different aspects machine learning to solve difficult mapping challenges.
- Engaged increasingly talented researchers to solve difficult geospatial problems. For example, one of the perennial SpaceNet winners (selim_sef) won the massive $1M Facebook Deepfake Detection Challenge.
- Brought together industry (Maxar, Planet, Capella) as well as academia (IEEE) and government (NGA) into a collaboration to advance the state of geospatial machine learning.
- Published eight academic papers by the SpaceNet partners, and 100+ academic papers cite the SpaceNet dataset.
- Gained massive traffic to the dataset:
Want to see more? Check out our video that further highlights these accomplishments:
We are honored to have charted the course for SpaceNet for the five years. SpaceNet resources may still be found at www.spacenet.ai, Twitter, and LinkedIn, and historical blogs will remain on the DownLinQ. As SpaceNet leaves CosmiQ Works’ stewardship, we are excited to see how academia, industry, and government will continue to use this unique, valuable resource to advance their missions.