Using satellite images to predict trends is Orbital’s unique offering. Read on to see how they do it, and where they are finding applications.
Sramana Mitra: Let’s start by introducing our audience to you as well as Orbital.
Jimi Crawford: I’m originally a PhD in Artificial Intelligence. I spent the first part of my career doing relatively basic research in Artificial Intelligence. I had the opportunity, about 15 years ago, to move here to Silicon Valley to lead a team in robotics at NASA Research Centre where we had fantastic projects in schedulers for the Mars rovers and genetically-engineered spacecraft antennas. Of course, being in the middle of Silicon Valley, I eventually got opportunities in startups that couldn’t be turned down.
I worked at a number of really interesting startups. In one job, it wasn’t a startup at all. I got an offer to run the Google Books project. That work, together with my work at NASA, is what led to Orbital Insight. One of the things I’ve observed throughout my career in Silicon Valley is that the opportunities for software guys in startups are very often driven by trends in the hardware. When the personal computer was invented, that gave the opportunity for Microsoft to exist. When the Internet was invented, it gave the opportunity for Google to exist.
The trend that we sit on top of is a tremendously rapid decrease in the price of satellites – the price of launching satellites and the price of the satellite hardware itself. The decrease in launch price is driven by what Elon Musk is doing. About six or seven other companies are building cheap booster rockets that are vastly cheaper than anything that has ever existed. Then the reductions in the price of the satellite hardware are being driven by the same trends that give you smartphones and the ability to build very powerful and small chips as well as cameras, radios, GPS. Basically, all the components that are in your cellphone are the same components that need to be on the spacecraft.
Our mission at Orbital Insight is to be the software layer of that trend – to build the cloud-based computing infrastructure that’s needed to take in a million satellite images, add them all up, figure out what they’re pictures of, count up all the things and images, and figure out on a global scale what they all mean.
Sramana Mitra: Can you take some of your customers and talk about specific use cases of how you address them.
Jimi Crawford: The very first product that we worked on after we founded the company was being able to predict corn deals. There are some years that the US produces so much corn that the price of corn goes down and farmers don’t make that much money because there’s so much corn out there that the price falls through the floor. In 2012, there was a relative drought in Illinois and the US corn production went down. We can look at the satellite imagery as well as humidity and rainfall data.
By combining them together pretty early in the growing season, we can get an amazingly accurate prediction of how corn will be produced. This is immediately useful for anybody who’s trading corn futures or anybody who builds a product for the government supply chain that depends upon corn syrup or the corn product so that they can start to hedge against the likely upcoming increase in the price. That’s the first part that we built. We’re now looking at extending that to look at other parts of the world.
That’s particularly useful outside the US because one, you can’t really buy crop insurance in other parts of the world. This early season prediction is part of what the insurance companies need in order to be able to start offering crop insurance around the world. It’s also a really useful early predictor of political instability because food security is often a major cause of instability around the world. We’ve seen, for example, that one of the proximal causes of the Arab Spring was the doubling of the price of wheat across the Arab World which traces back to a famine in the Ukraine. We can also give early warning of possible food security issues. That’s one big use case around satellite imagery at scale to track global agriculture.
This segment is part 1 in the series : Thought Leaders in Artificial Intelligence: Jimi Crawford, CEO of Orbital Insights