With every piece of equipment and every line of computer software that the precision ag sector creates, farmers are playing catch-up. Yes, it’s exciting to learn about new technology and the endless possibilities around the corner, but in the real world, that corner is still out of sight. The requirements for huge amounts of data and the ability to respond to that data are still in dreamland.
Maybe, though, we’re seeing the shape of a solution.
The latest piece of precision ag equipment to launch — literally — is nanosatellite technology. An incredible leap forward in the design and deployment of orbital platforms, nanosatellites are considerably smaller, lighter and quicker to develop than the satellites we’ve always read about. They’re less expensive to build and launch, too, and they have lower trajectories and can orbit the globe 14 to 16 times per day.
Conventional satellites with sizes that range from small cars to cub vans can take five to 15 years to design and build, and costs can run as high as $500 million.
That’s according to the website of Alén Space, a company based in Spain that has been developing nanosatellites since 2007, adhering to CubeSat standards.
Nanosatellites weigh between 2.5 and 25 pounds, have a volume of one to 10 litres, and cost as little as $500,000 to develop and deploy, based on an eight-month construction process.
The United Nations Office for Outer Space Affairs (UNOOSA), states that more than 8,000 objects have been launched from Earth since 1957. Partly due to nanosatellites, it now expects the number of such objects to, well, take off.
Early in 2020, a news release described a study conducted at the University of Illinois, comparing how nitrogen stress in corn was measured by unmanned aerial vehicles (UAVs or drones) versus CubeSat-based multispectral sensing technology. The researchers, led by Kaiyu Guan, Yaping Cai and Emerson Nafziger, had monitored nutrient stress in 28 nitrogen (N) management treatments in central Illinois in 2017.
Drones and CubeSats were set up to detect changes in corn chlorophyll concentrations due to varying N application rates and timings, and the research found that two technologies worked with a similar degree of accuracy.
What the study illustrates is not only the capabilities of nanosatellites but the technology’s vast potential. According to the university’s news release, there are now more than 100 nanosatellites in orbit, with more on the way.
There are still issues, including familiar precision ag issues like too much technology, questionable cost-efficiency and the extraordinary amounts of data generated.
Still, in discussions surrounding the use of any imaging technology, it seems costs are only a short-term hurdle. Like computers 25 to 30 years ago, prices come down as usage increases, although there’s disagreement over whether there are enough farmers to “lighten the load” quickly enough.
Rising on-farm costs are challenging most producers and an expanding set of variables are being factored into most equations involving investments in new equipment.
But hidden in the talk of potential opportunities is the one basic question that has to be asked of any precision ag tool: What do you want to do with it?
For images, it’s not a question of whether a grower is using drone technology or satellites — or nanosatellites — that’s important. It’s understanding the goal. For example, the Sentinel satellite now flies over parts of Ontario four times a week, using resolution that’s publicly available. If a grower or retailer or agronomist wants higher resolutions, someone has to be willing to pay for it.
“We also have to make sure we’re comparing apples to apples,” says Brandon Yott, strategy and business development manager with A&L Canada Laboratories, in London, Ont. “It’s hard to compare drone to drone imagery, and that includes discussions on high-resolution RBG (red-blue-green) cameras versus multispectral versus hyperspectral cameras. If I can read a license plate, that’s the resolution of an RBG camera, like the ones on a phone. But resolution can be a function of the camera, it can be a function of how many pictures you take for overlap and stitching images together. It can be a function of how high and how fast you fly.”
For something as simple as a contractor installing tile, an RBG camera will do the job. But Yott adds that when it comes to reading soil nitrogen or CECs, predicting yield or detecting stress, an RBG camera is insufficient.
This is where costs become a key consideration. Lower-end multispec cameras, which capture a few wavelength bands can cost as little as $2,000. Higher-end multispec cameras, capable of capturing up to 10 bands, are closer to $10,000. That’s still relatively inexpensive next to hyperspectral units that capture the entire bandwidth and are more for military and research uses: those can run in excess of $100,000.
“But the bigger barrier is not the dollars, it’s the dataset,” says Yott. “Let’s say I’m flying over a 10-acre field and I can capture decent resolution with an RBG camera at two to 20 MB. If I want to do that with our multispectral camera where I have seven bands and an RBG that’s not super high res, I’m at two to 20 GB. If I’m going up to that hyperspectral, I’m likely up around 200 GB.”
The computers, the storage and cloud capacity required to deal with that amount of data is almost as astronomical as the potential for the technology. The good news is if a grower simply wants to determine when to apply nitrogen, a standard NDVI, with additional bands and RBG can provide sufficient resolution. (There is some new research looking into N recommendations with only RBG imagery, where they’re using a newly developed Greenness Index to look at the colour of leaves, chlorophyll and nitrogen relationships.)
Benefits and drawbacks
Yott concedes that both drones and satellites have their strengths and weaknesses. Satellites — and, yes, nanosatellites have the same challenge — offer imaging with reasonably high resolution that is publicly available, with frequent orbits on a weekly basis. But cloud cover can render those images unusable, even with four weekly passes over a region.
“If I’m two days away from doing my nitrogen application and I need to know, and I know I can do it with the drone or I have to cross my fingers for the satellite, I’m probably pulling the trigger on the drone,” says Yott.
At the same time, drones are more labour-intensive and more costly in their operation. Also, the frequency of flying over a field isn’t the same as with a satellite. Yott likens it to the discussions surrounding robotic weed management: as advanced as the systems have become, growers’ knowledge surrounding the use of the systems isn’t keeping pace.
For Mike Wilson, that steep learning curve is one of the bigger impediments with precision ag adoption, including satellite versus drone imagery. Nanosatellites are a wonderful evolution but it’s technology that is largely unusable by the vast majority of producers.
That isn’t meant as a criticism as much as an acknowledgement of the demands of farming. Shrinking margins and rising costs are stressing management decisions, and the time needed to study new technological systems and implement them can simply be unrealistic.
“Growers are embracing technologies that have been tested and proven,” says Wilson, affiliate program lead and a certified crop advisor (CCA) with Veritas in Chatham, Ont. “It takes so long to see results that it’s hard for growers to keep up with the pace that technologies are changing. For certain, a drone image is centimetre-quality resolution versus three-metre, and you can see a lot more with a drone than a satellite. But if you’re going to pump that data into a 60-foot-wide drop boom, three metres is good enough.”
If a grower has electric drives on a planter and the farm is variable and there’s individual row population control, then a higher, sub-metre resolution is necessary. But even then, the basic question always comes back to “What do you want to do?”
“Until we get to a cost-effective source of that technology, I can’t see the majority of farmers implementing it,” says Wilson, stressing that he’s not trying to be critical, just recognizing the demands on the farmer. “Until there is easy-to-use software that provides actionable results from these large datasets and the technology in our equipment allows for more precision in our application, these large datasets will struggle to show their real value and farmers’ adoption rates will be slow.”
But again, that’s the drawback on the technology side in agriculture. Unlike computer software or cell phones — which evolve rapidly and discard older models — agriculture is very selective about what it declares obsolete. Varieties and hybrids are discontinued but farmers still use atrazine or older formulations containing dicamba. As more new systems, designs and chemical products are introduced, they’re added to the knowledge base. That means as fast as developers and manufacturers can innovate, farmers are forced to either expand their understanding or hire someone to learn it for them.
Wilson concedes that’s an imposing challenge for many, particularly since cell phones and computers can see the advent of new systems, implement them and test their viability and move forward with them, while in farming, the learning and evaluation process can often take 18 months or even considerably longer.”
“You decide now whether you want to make a change in your operation, so you try it this growing season, you get the results back and if you’re lucky, you would have something corresponding to next year — probably longer,” says Wilson. “By the time you figure out if something works, you have a whole year that’s gone by and everything’s changed — it’s all new again.”