Interview: Earth-Imagery Automation Via Flexible API

by | Dec 5, 2015

bio_headshotAstro Digital is working to simplify how users discover and receive satellite imagery. The company has already built an online platform with a flexible API to improve delivery of Landsat imagery and other data. The platform will also be handling all the imagery from its smallsat partner Aquila Space, which plans to launch a fleet of Earth-observation satellites. Earth Imaging Journal's Editorial Director Matt Ball spoke with Bronwyn Agrios, co-founder of Astro Digital, about the company's creation of an online image-processing and distribution platform to send real-time alerts when satellite imagery is collected across an area to be monitored.

EIJ: How long have you been working on the Astro Digital satellite-imagery platform?

Agrios: We started Astro Digital at the beginning of 2015 to fill a demand in the market to give developers easy access to data from space. A lot of our initial motivation for starting the company was to release software products that make it easier to work with satellite imagery and at a quick pace.

Historically, you have needed to be a professional user of GIS or remote sensing and have a lot of training in the tools to actually work with Earth-imaging data. In my experience working in the geospatial industry for some time, there's a real barrier to entry for those who don't have the expertise. This is true of maps in general, but particularly with remote-sensing data, because there's a large level of complexity in satellite images.

EIJ: In the history of new satellite imagery, the satellite usually comes before the means of delivery. What made you prioritize data delivery before having satellite imagery from your Earth-observation partners?

Agrios: We chose to invest heavily in the product development up front so that we could begin immediately offering useful tools for working with satellite data. Our focus has been on monitoring workflows and thanks to Landsat data, we didn't have to make a huge investment in commercial data upfront. I couldn't be happier about how the product is coming together using an agile approach and open data to start vetting our ideas and building useful product.

We originally released a search API, again because it's not easy to browse, preview and integrate satellite data. There wasn't an API out there for developers to use or even for that simple search-and-discovery operation. An awesome outcome is that we have helped to make Landsat more open by making it accessible. Ultimately, we are moving toward more of a monitoring platform, which we just released in the last couple weeks.

EIJ: Who is your customer?

Agrios: Anyone with basic web development skills can start using the API. We've also got a great interface at to search, process, and publish maps to the web. In the commercial satellite space, everyone is looking for the potential size of the remote-sensing market. Our focus is primarily on land use monitoring, which is a sector that isn't being fulfilled by traditional satellite imaging operations and business models. By remaining low-cost and agile, we're testing workflows and business models to prove where some of those commercial opportunities lie.

EIJ: Do you think the democratization of data and more-agile tools are what it will take to break open the commercial market?

Agrios: Yes, better tools will lead to more access. The business and access models have really been fine-tuned to serve government. The reality is that data isn't accessible, and the business models don't make it affordable enough for end users to get enough data to actually impact the way they do business.

There are obviously a few things driving the changes we're seeing. There is the advent of more monitoring constellations coming online. The Landmapper constellation from our partner Aquila Space, Planet Labs has big hopes for daily imaging of the whole world, and companies like Urthecast and BlackSky are all getting into the game with high temporal resolution constellations.

That increase in data is creating a different paradigm for imagery use. Today we point and shoot, where we say what area we are interested in, and we pay for the whole picture. That business model will remain, but with the monitoring concept we are going to have a huge amount of data dumping down from the sky. We need to have the tools to make sense of all that data.

The platform we're building at Astro Digital is meant to help developers and consumers of satellite imagery optimize their workflows to take advantage of those monitoring constellations.

EIJ: What's different about the monitoring mindset?

Agrios: With always-on monitoring constellations, users don't have the capacity to pull in all the data all the time. It's just not practical. We want to optimize the way people get the data and to only be sent the data that is interesting. Today, we're data poor. In the very near future, if the investment in small satellite constellations pans out, we're going to be very data rich.

We need to know what to do with those data sources. We've created the means to define an area of interest, to set up filters that determine what changes you're interested in then define when and how you want data sent to you.

For instance, if you manage farm fields across the country, our platform can monitor every field for new information by assessing new imagery for cloud cover quality, analyzing the pixels in real time to compare against the values for crop health from the previous season, then send you a map plus values for only the fields that exhibit significant change.

EIJ: Our ability to detect change hasn't been great in the past. Will automation make such insight more accessible?

Agrios: Automated change detection is difficult. We're pretty deep into the research and development, and we're starting to see really good results. I'm excited to start rolling it out into the product.

EIJ: What is your relationship with Aquila Space, and what is the timeline for imagery from those satellites?

Agrios: Aquila and Astro Digital are good friends and we are now officially a reseller for the Landmapper constellation. Aquila has a talented team for designing, building and operating small satellites. They have designed impressive spacecraft for this mission. One of the amazing capabiliites is a Ka-band downlink that gets data down to Earth much faster, which is a huge bottleneck.

Aquila's Landmapper constellation will start with Landmapper BC, which is a 22-meter constellation that will cover huge amounts of the world with a focus on living resources. That will be complemented by the Landmapper HD constellation, which is 2.5-meter resolution. The two together will provide a really nice ability to understand the broader context to see what's going on and then be able to drill down into more detail with the 2.5-meter satellites.

We're working very closely with them to anticipate the use cases for the monitoring satellites they're building. When those satellites become fully operational, we can flip the switch and get data out into the market as soon as possible.

EIJ: Is agriculture your biggest potential market? Where are you seeing the most traction?

Agrios: We're definitely seeing the most traction in agriculture, as they're really starved for Earth-imaging data right now. The tasking model is not conducive to large-area coverage or monitoring.

There is a lack of reliable repeat coverage from the current satellite operations model, but once we have a lot of data coming from space, then we're going to have the ubiquity needed to be reliable. For companies that already are used to working with sensor data, such as in agriculture where they rely on weather data and data coming from their equipment, we need to move Earth-imaging data into that same direction to improve adoption.

EIJ: Do you plan to expand the imagery sources offered?

Agrios: Yes. Without getting into specifics, we are working to plug into all the satellites”whether commercial or country-level satellites. We're like the intelligence layer that lets people control the monitoring for themselves. That's going to require a lot of data, so being able to provide that intelligence layer, which takes users away from the complexity of the sensor systems themselves, means we will have the ability to work with a lot of different companies. And we will need to work with a lot of different companies to deliver insight into change.


May Issue 2024