- October 30, 2017
- Posted by: EARSC
- Category: Members News
From October 9-11 more than 150 geospatial and remote sensing professionals gathered to discuss the groundbreaking ways in which optical satellite imagery is being used across a wide variety of applications.
Automated processing of big data, using optical imagery to make predictions about the future, and the potential of new image collection strategies were recurring topics of discussion, proving how far the industry has progressed since European Space Imaging was established 15 years ago.
“We are really at the edge of virtualizing good parts of our planet and life,” said Dr. Stefan Voigt, Senior Researcher at the DLR German Aerospace Centre. “We are adding a whole new sphere – the digital sphere – and that makes things ever more complex, the challenge is more and more to understand the impact of this virtualization and of our digital footprint.”
The first big-news item on the agenda was the recent acquisition of WorldView Global Alliance partner DigitalGlobe by MDA Corporation, and the subsequent launch of a new parent company named Maxar Technologies.
“Maxar gives us an incredible competitive advantage moving forward,” said DigitalGlobe’s Andrea Bersan. “It will allow us to continue to do what we do, and we are going to do more and better.”
Mr. Bersan presented the company’s plans for Scout and Legion – DigitalGlobe’s new satellites that will almost double the coverage capacity of the constellation, bringing it up to 6 million km2per day. Attendees were especially interested in the technical specifications of the new satellites.
“We’re building a constellation to image the most interesting parts of the world most often,” said Deke Young, Senior Manager of Product Development at DigitalGlobe. “We’re also establishing a repeatable collection strategy that our clients can depend on.”
High level representatives of the European Commission and EU agencies outlined current and future European programs and how they use VHR satellite data for decision-making and operational purposes.
“Particularly in emergency situations, this data can prove to be very valuable information,” said Sónia Antunes from the European Maritime Safety Agency.
“High-revisit data in location-based apps will open up new possibilities in the consumer market,” commented Steve Bochinger from Euroconsult during his presentation on market trends. “Especially when integrated with AI and machine learning.”
There were some innovative use cases of automation and deep learning, with two speakers discussing projects that are using satellite imagery to make predictions about the future with a high level of accuracy.
Spandan Kar of Space Imaging Middle East presented his work for the Smart Government Initiative for the Dubai Police. By analyzing geospatial information relevant to the city’s crime, they are enabling optimization of police force operations, and ultimately aiming to predict the activities of the city’s criminals.
“Out of every 100 crimes that happened, so far we have predicted 52,” Mr. Kar said.
Satellite imagery is also being used for its predictive ability in the field of natural disasters. Hinnerk Gildoff from SAP HANA presented a case using deep learning algorithms to predict the locations where wildfires are likely to occur.
“If the risk of a fire is higher than 80% we give out a warning,” he said, “so far our estimates show an accuracy of more than 85% based on our test data, but we want to reach 90% before we deliver it.”
Mr. Gildoff hopes that the program will help fire management authorities to focus their attention on the right geographic areas, and ultimately save lives.
During the session on humanitarian mapping, speakers from multiple relief organizations highlighted the value of satellite imagery when making decisions about the most effective way to respond to crises.
“There is a huge need and demand for information about conflict affected countries and satellite imagery allows us to access it from a distance,” said Sarah Muir from the United Nations’ World Food Programme. “What we are in need of are automated techniques to improve the efficiency of the process.”
Other speakers discussed projects in which they are developing cutting edge image processing techniques to respond to the needs of end users. Of particular interest was Tobias Leichtle’s work for the Company for Remote Sensing and Environmental Research (SLU), in which he has developed an approach that automatically detects changes in the number of buildings in a given area.
“The spatial resolution of LandSat data was not sufficient,” Mr. Leichtle said. “This approach gives us a highly automated way to get information about the relative dynamism of different cities.”
With European Space Imaging’s state-of-the-art ground station collecting more data from the WorldView constellation than ever before, and with a human-driven image collection strategy that minimizes cloud cover, there can be no doubt that it is poised to respond to stringent requests from innovative government and business customers.