The Growing Ecosystem of Satellite Imagery for Ag

There has been a lot of hype around the use of satellite imagery for agriculture. The idea that images taken from above can help farmers understand what’s happening in their fields is exciting, and companies working across industries to image the planet every day are raising big bucks from the venture capital community.

Images taken with a simple RGB filter can help farmers and their service providers, such as agronomists or input companies, to map out and prioritize zones in their fields that need attention based on how green they look compared to others. Planet Labs’ newly acquired RapidEye satellites go one step further into measuring crop stress, using near infra-red filters (NIR) to measure colors the human eye cannot see. NIR images are usually analyzed through the normalised difference vegetation index (NDVI), which is a simple graphical indicator measuring the amount of live vegetation in a plant, such as its chlorophyll content and biomass.

These uses of satellite imagery are certainly useful, and some companies have built successful businesses from them such as Geosys, the 27-year old satellite imagery company that was acquired by agriculture cooperative and food brand Land O’ Lakes in 2013. Geosys sources daily images from up to 10 different satellite companies to build its own virtual constellation and ensure it has enough images of the right resolution to serve its clients. Geosys uses NDVI to process maps for its clients the next day. Geosys’ clients are typically not farmers but the service providers working with them like seed companies or advisors.

Agricultural operators are not the only potential clients for satellite imagery companies; there’s a growing business around providing commodity traders and insurance companies with data on predicted yields and so on.

As the client-base expands, and the limitations of current offerings come to the fore — limitations such as the need to ground-truth so much of the data coming off the satellite imagery or the infrequency of relevant images means events are missed — a growing ecosystem of startups is coming to the fore, promising to offer clients actionable insights.

From startups destined to launch satellites into space themselves to software analysis tools promising to process images faster than anyone and give farmers specific instruction that day, these startups are all approaching the challenge and opportunity for satellite imagery in ag in a slightly different way.

“The first step is making the images available — that’s what we’re really good at. And now there’s an ecosystem of businesses popping up to analyze the huge levels of data that Planet Labs is putting together in one location,” said Ryan Schacht, director at Planet Labs. “We are definitely seeing a lot of growth in the number of companies that are realizing this level of data didn’t exist before which are now building a business analyzing it. This network of providers have the domain-specific expertise to answer questions on behalf of the industry. They can apply contextual information for analysis to turn data into answers.”

Here’s how three new startups approaching satellite imagery for agriculture industry is slightly different ways.

FarmShots, a startup from North Carolina, wants to give farmers actionable insights around disease management from satellite imagery.

“It’s often just a pretty picture for the farmer,” says Joshua Miller, CEO of FarmShots. “It’s telling them there might be a problem in one area of the field, but it’s not telling them what it is. And by the time the farmer has verified the issue and gone back and forth to the field before applying it, he’s missed an opportunity.”

Through its large annotated databases of certain crop diseases, FarmShots is aiming to automatically detect a specific disease on a farm through satellite imagery. By comparing new satellite images that may contain plant diseases to previous images that are confirmed to have diseases, FarmShots aims to extract the specific disease that is affecting the field.

Farmers are then alerted with a PDF report that details the precise problem and the solution to be applied in real-time. It has an app and costs $1 per acre for a yearly subscription.

Miller built FarmShots out of Duke University, where he graduated two months ago. He built the company on the revenues from early clients and is only just going to market to raise a seed round. From the first iteration of software Josh built himself three years ago, FarmShots now services 300,000 acres of land. “I’m not a great engineer, it’s less about building some dazzling piece of software, and more about listening to the market and building what customers want,” he said. “We’re at the end of a super cycle in commodity prices. If you’re not providing value to farmers, farmers do not have enough of a margin to adopt your technology unless the value is really obvious.”

Astro Digital is building a business-to-business platform offering imagery analysis from its own constellation of satellites to farm management software companies, commodity traders, insurance platforms, weather applications, and any other clients monitoring global agriculture.

It is planning to launch two different types of satellites; one with red, green, and NIR filters to monitor vegetation growth on a daily basis for production agriculture clients, and the other with RGB, red-edge, and NIR filters to get early detection of crop deterioration

“We would love to do short-wave infra-red too and thermal imaging which can help with soil moisture measurements, but we’ve got to keep focused,”said Bronwyn Agios, co-founder of Astro Digital.

It will launch its satellites later this year to target specific regions, but in the meantime has built a software and imagery processing platform using data from other satellites such as Landset.

“We’re building a system for monitoring global agriculture; we’re selling information, not pixels,” said Agios, who used to work at Planet Labs. “We are the only satellite company that’s building software for analyzing and distributing satellite imagery data at the same time as building hardware.”

This means that Astro Digital already has clients plugged into its API that are accessing processed imagery from other constellations. “By building our own hardware we will have the ability to offer our clients better resolution, more frequent imaging, and more reliability through the software platform they’ve already integrated with,” she said. “You tell us where you want us to monitor and we will deliver real-time data to you through our API.”

Vinsight, a Redwood City, California startup, is targeting high value cash crops with an initial focus on grapes, and aiming to give the industry accurate yield predictions, a pain point for the industry, according to CEO and co-founder Megan Nunes. Nunes used to work at Canopus Systems, a small-satellite company affiliated with Dauria Aerospace, alongside her co-founder Tomas Svitek.

“Yield prediction is critical and has traditionally been done by counting berry clusters and extrapolating out the numbers by acre size, but because grape growers are in micro climates, yields will vary in loads,” she said. “So we thought ‘how can we make that number 30 percent to 40 percent better?’”

Nunes now promises clients a 5 percent to 10 percent error rate four months before harvest, using satellite imagery, climate, weather and historical data. “A picture doesn’t equal actionable intelligence,” she said. “In order for satellite imagery to become something fully adopted by the agri industry, there needs to be further actionable insights on that imagery. By offering yield predictions to such an improved degree of accuracy, we are offering that value-add.”

Vinsight uses public satellite imagery data to measure the vegetation of an area and then compare that data with 10 years of weather records and harvest reports to detect patterns and identify correlations.

Farmers pay on a per acre basis for a subscription to Vinsight.

Have news or tips? Email [email protected]