Maxar continues to engage the SpaceNet partners to plan the next phase of the project and future challenges. A collaboration between DigitalGlobe, CosmiQ Works, and NVIDIA, the SpaceNet challenge leverages an online repository of publicly. US: DigitalGlobe has announced the SpaceNet Challenge Round 2 results and plans for the next two SpaceNet Challenges. Maxar believes that AI enables huge benefits for the speed and scale at which we can serve important missions, such as rapid mapping after natural disasters, maintaining updated maps globally and understanding our changing planet.” The SpaceNet Challenge Round 2 results has been announced for the next two SpaceNet Challenges. “We are grateful for IQT Labs’ outstanding leadership of SpaceNet, and Maxar is thrilled to have the opportunity to lead the next phase of innovation. ”SpaceNet has become the gold standard for geospatial computer vision reference datasets and has fostered open innovation for algorithms to automate the extraction of foundation mapping features,” said Tony Frazier, Executive Vice President of Global Field Operations at Maxar. During this planning period, rest assured that all existing SpaceNet resources, including training datasets, tools and algorithms, will remain available. We’re excited to explore the application of AI to new sources of imagery, 3D data and other sources of geospatial data that can help us better understand our planet.
We are pleased to share that we plan to host SpaceNet 8 in the second half of 2021 we will share details when possible. We look forward to continuing SpaceNet’s impressive track record of open innovation as the geospatial community takes on new mission challenges. Maxar commends the significant contributions of the CosmiQ Works team for managing this project for the past five years, and we’re grateful for the collaboration. Under CosmiQ Works' leadership, SpaceNet completed seven challenges, developed 36 open-sourced algorithms, had participation from more than 80 countries and released about 67,000 sq km of high-resolution satellite imagery, from which 11 million labeled building footprints and 20,000 km of road labels were mapped. SpaceNet is dedicated to accelerating open-source, AI-applied research for geospatial applications, specifically foundational mapping (i.e., building footprint and road network detection). Today, we are excited to announce a leadership transition of the SpaceNet project to Maxar. Maxar co-founded SpaceNet with IQT Labs’ CosmiQ Works five years ago to harness open-source innovations in emerging computer vision technology to accelerate complex analysis of satellite imagery. In this paper we discuss the SpaceNet imagery, labels, evaluation metrics, prize challenge results to date, and future plans for the SpaceNet challenge series.From smart speakers to facial recognition, artificial intelligence (AI) is increasingly part of our daily lives.
The first two of these competitions focused on automated building footprint extraction, and the most recent challenge focused on road network extraction.
The SpaceNet partners also launched a series of public prize competitions to encourage improvement of remote sensing machine learning algorithms. Accordingly, the SpaceNet partners (CosmiQ Works, Radiant Solutions, and NVIDIA), released a large corpus of labeled satellite imagery on Amazon Web Services (AWS) called SpaceNet.
We propose that the frequent revisits of earth imaging satellite constellations may accelerate existing efforts to quickly update foundational maps when combined with advanced machine learning techniques.
Updating maps is currently a highly manual process requiring a large number of human labelers to either create features or rigorously validate automated outputs. Foundational mapping remains a challenge in many parts of the world, particularly in dynamic scenarios such as natural disasters when timely updates are critical.