The data analytics industry has come a long way in its career, turning images into meaningful data for decision makers, but there is still work to be done to make machines better able to extract information from data without a process, Scot Currie, Chief Solutions Architect at Earth BlackSky Observation Company predicted.
The current problem is that Artificial Intelligence (AI) and machine learning are not mature enough to fully automate analysis, Currie told SpaceNews on the sidelines of the GEOINT Symposium 2021.
The element of speed is how fast we can make machines do what humans have been doing for the past 50 years? said Currie, who recently joined BlackSky and previously directed imaging programs at the National Geospatial Intelligence Agency.
BlackSky, based in Herndon, Virginia and Seattle, defines itself as a “global surveillance” company that uses multiple types of data to detect changes or anomalies. It uses machine learning algorithms to analyze optical and radar images from satellites, images collected from airplanes and drones, data from ground sensors, Internet of Things networks and social media.
The recently listed company operates six satellites that fly in inclined orbits to focus coverage on global hot spots and target a constellation of 30 satellites in the coming years. Most of its activities currently come from government agencies.
The challenge today is that customers have to trust the information, and that still takes a lot of work from human analysts, Currie said. Industry is pushing for technology to advance, but surveillance is not fully automated yet.
Currie predicts that machine learning will predominate users of pixels in the next 10 to 20 years and that analysis will place fewer demands on people.
This will be a race against time as analysts are faced with a data overload. The pictures are getting bigger, said Currie. Analysts get overwhelmed with images, so you have to go in a different direction. You need to make the machine count, determine that something is going on, and give you a warning. Then you can assign a collection item for review.
BlackSky and other industry players are working to mature the algorithms to train them to be accurate enough with the resolution we have in our constellations, he said . Will there be false positives? Yes, but people make mistakes too.
