Sensors & Systems (S&S) editor Matt Ball spoke with Dirk Robinson, vice president of Image Systems at Skybox, about the company’s progress, the focus on an online data platform, and some of the early responses to their new imagery and video from SkySat-1.
S&S: I noticed from your online profile that you have worked at a number of companies with varying jobs all associated with digital imagery. How long have you been with Skybox, and how are you applying your knowledge of digital imagery?
Robinson: I’ve been working here for about four years. I was here when there were twelve of us in a small thousand-square-foot office, drawing out designs on napkins.
Imaging is actually pretty tough. In this day and age we have SLRs that produce really stunning image quality, and even our Apple iPhones continue to produce better imagery, but that’s really the result of decades of development.
Taking pictures from space is no different, and imaging satellites are one of the more complex systems to build. There’s a whole host of things that can go wrong. At the end of the day anybody can look at a picture and say, “You know what? There’s something wrong with this,” or “You know, this is a really beautiful picture and I can learn a lot from it.”
I’ve been doing imagery for more years than I care to reflect, but my interest and passion has always been in creating really high-quality sensors, as well as extracting information out of imagery. I’ve gone back and forth through my career in designing novel imaging systems and trying to gain insight from imagery.
At the core is the interplay between what a sensor is, how it captures information about the world, and also how to do that efficiently. In order to do that, you have to have a deep understanding of how the algorithms work and how the measurements you take about the world can be used to answer a question.
Before I joined Skybox I was doing that for other industries, working on automotive, medical and consumer imaging applications.Here at Skybox I find there’s really no difference. We’ve tried to take a vertically-integrated design approach—a combination of both the spacecraft that we’ve built, which is much lower cost, and marrying that with sophisticated ground-processing algorithms, in order to create really high quality pictures. From there, we then have the means to extract a phenomenal amount of information.
S&S: Just recently there was the announcement that you’ve signed a contract to build thirteen more satellites with Space Systems Lora. That must be exciting to be ramping up the planned 24-satellite constellation.
Robinson: It’s exciting to be working with Space Systems Loral (SSL). They have a rich history in the space industry, where there’s no room for failure. Their decades of experience and a demonstrated track record of success gives us a lot of confidence that when we go out and build our constellation of satellites that they’re going to be built to the highest quality standard.
S&S: Are there advantages of having another California-based company like Skybox building the satellites?
Robinson: That’s icing on the cake. They’re a few miles away, so it makes it really easy to work closely with their team. While we’ve done a lot of design work to make our systems easy to manufacture with a standard repeatable process, there is always a lot of communication and information transfer that has to happen between our two companies. It’s so much easier when they’re right up the street. We have desks for them and they’re going to have desks for us. It’s much more convenient than having to get on an airplane or having to Skype in order to work with your partners.
S&S: I recall that you created your own clean room as part of the operations, and were doing your own manufacturing. Have you moved past that mode to scale your operations with a larger satellite manufacturing company?
Robinson: We will always retain the capability for designing next-generation satellites, and their continuing technical evolution is a core part of our business DNA. We’re going to retain that design and prototype capability here at Skybox headquarters.
In the satellite world, as you know, high volume is not in the thousands of units. Even single dozens of units is something that’s rare. It’s even rarer to do that with the performance that SSL has done.
S&S: Congratulations on the Skysat-1 satellite launch success. We’ve seen some imagery, and even some video. Is there ongoing work with that input to create data products and services?
Robinson: Yes. We have three basic product lines; we sell pictures, we sell an access product, and a third business line is the information product, which is information services around insight derived from the imagery. We’re working closely with a small number of customers and partners as we develop that business line. We’re also busy focused on continuing to get our first satellite operational and producing high quality data day in and day out. This is a three-shift, seven-day-a-week global operation. We’re still in the calibration and validation phase, but we hope to be transitioning out of that into the operational business phase in the next few months.
S&S: Originally you had plans to be operational in December of 2012, and I know you ran into scheduling difficulties with the satellite launch in Russia. There’s a great deal of complexity to this business. Did the delay cause problems?
Robinson: The information on delays was a periodic process with changes every few months. It was frustrating because we put so much heart and soul into the first two spacecraft that it was like having your wedding day shift every couple months. Looking back, I think it really allowed us time to do a couple things: it allowed us time to build a second satellite, which we weren’t originally intending to do, and it gave us the opportunity to prepare more for operations. I actually think that we are moving very swiftly through our early check-out calibrations. And that’s due to the extra time that we had to prepare.
S&S: Your sensor has a unique approach of capturing both video and imagery. What was the evolution of this offering? Is the capture of video imagery simply a capability based on the sensor configuration or algorithms and processing?
Robinson: It’s a combination of both. We do have a unique capability in our imaging system, called a two-dimensional imaging sensor. This is the same type of imaging sensor that you’d have in your phone or your DSLR. It’s basically a 2D chip that takes entire snapshots at a single point in time.
That’s in contrast to most other high-resolution earth-imaging satellites that have meter or centimeter resolution. They use what’s called a TDI or time-delay integration imaging system. It’s kind of like a copy machine; it’s a 1D system that scans over the earth, and you can create very high quality pictures but it requires you to have very expensive pointing control on your spacecraft.
It’s another Skybox story about leverage, and in this case there is a new growing market in imaging sensors where you can start to have very high resolution, very high performance, very high frame rate in two-dimensional imaging sensors. They are fairly new in the world of imaging, and we’ve decided to take advantage of that and build our entire imaging architecture around it. The two-dimensional imaging sensor also allows us to do video. That’s something that you can’t do when you only have a one-dimensional imaging sensor.
It also allows us to push a lot of our complexity into our ground processing. So, for example, with a one-dimensional system you have to maintain the stability of your spacecraft; it has to be pointing somewhere with extreme precision and stability. In our case, we can allow it to wobble a little bit, so to speak, at which point we’ll take that data on the ground and use very sophisticated processing algorithms to essentially clean up the data and create high quality products. Wherever possible, we’ve tried to push the complexity out of the spacecraft and into ground processing, where we take advantage of Moore’s Law. We’re always going to have more Computrons at our disposal whereas putting it in space just drives up cost exponentially.
S&S: The video capture is unique, and very interesting. Have you seen a good interest in that capacity?
Robinson: We have, although it’s still very early. We are the first commercial provider of high-definition video from space, and we are working with a lot of our early customers. One advantage of Skybox is we can actually start a business and make a fair amount of revenue with one satellite in a still imagery market, which is the market of today. And so the video product is very exciting.
There are a lot of folks who are eager to get this. We’re still working out how to position this in the marketplace, including how to sell it and who’s going to derive the most benefit. We’re actually releasing a new video sequence today on our website that shows some of the interesting things you can do with video. There’s one video of a coastline and a coastal town and you can see the waves breaking. We’re working with a lot of partners who are very interested in things like understanding wave dynamics to determine what the surface of the seabed is. There are a lot of people who want to understand what the topography under the sea floor is, and you can only capture the wave dynamics from satellites when you have video.
Having video is something that’s unprecedented. We’re working closely with partners, but again, I think it’s still pretty early in the market for high definition video to discuss all the market or all the possible uses.
S&S: One thing that I want to touch on is the Internet DNA within the organization. Is the idea to feed web services with your imagery, and to feed information down to mobile even, to make this data accessible anytime anywhere?
Robinson: I see this as one of the big stories of Skybox. We’re a really vertically-integrated company, but, in the same way that we take advantage of highly reliable automotive parts in building our satellites, we really want to take advantage of all the information technology development in Silicon Valley.
The amount of data we’re getting is really breath taking, the enormous amount of data. You have to be just as smart in designing your information technology so you can bring all that information together to find the data that you want. When you have petabytes of data you have to make sure you can bring all this information together, to run that analysis, to generate that report, and then be able to deliver information out to customers.
As we move forward, we invest a lot of time in figuring out how to build a platform with accessibility for other developers to bring their algorithms onto our platform. That’s another core part of Skybox: it’s not just the accessibility of image data, but the accessibility to run your analysis and take advantage of your own know-how.
We know that we’re not going to employ all the data analysts, and it would be a silly idea for us to do so. In this space there really are pockets of people that have domain expertise, such as the world’s experts on oil pipelines that have either an automated algorithm or a very simple work flow for monitoring oil pipelines and seeing if there are backhoes parked next to it (which might indicate a pending problem).
We need access to data and information, but also a place where we can run analysis. Down the road you have a place for individuals to monetize that with a marketplace. We spend a lot of time thinking about that – in the same way that Orbitz revolutionized the airline industry and took a lot of the manual labor out of the process, we’d like to do the same thing for the remote-sensing and space-based information business.
S&S: I really like the company tagline, “analytics of daily global activities.” It’s all about teasing out these little science and analysis-based discoveries, would you say?
Robinson: At the macro-scale I believe what we’re trying to do is transition an industry. The commercial satellite imagery market has been underway, but now we’re moving out of the realm of big government and big well-heeled institutions and into the commercial space.
That’s with regard to our satellites, but it’s also about understanding what this data means. In the market today, if you want to get a picture you’re going to get a big two-gigabyte or larger image file. You’re going to need a $10,000 piece of software to load it up and you’re probably going to want to hire someone with a master’s degree to understand what to do with it.
A large part of the Skybox vision is to simplify that process to get access to the information underneath. At the core of a lot of these analysis and information products is sophisticated science, but we’re trying to provide access to that information in a way that makes it very easy for a CEO or anyone who’s got to make a business decision. Questions like, “Do my crops need more nitrogen?” or “Which one of my fields still has a particular visible parasite?” I know these users don’t want to have to hire a PhD or data scientist. They’d like to just have a daily or weekly report. That’s really where we’re trying to take the marketplace.
We’re excited to let the world know about what we are doing. We’re really about growing an entire marketplace and so we’re excited as other folks are coming online with new data sources. We love playing with data and we believe that this is really going to transform the way businesses and humans understand what’s going on, on a global scale.