The Augmenta Field Analyzer on CNH’s sprayer uses CVML and AI to translate camera images into agronomic insights, fine-tune prescriptions, and implement variable rate applications in real-time; the capability to identify weeds is in development. | Source: Brianna Wessling
Just one U.S. farm, on average, can feed 166 people annually in the U.S. and abroad, according to the Farm Bureau. Every single day, you eat a fruit or vegetable that got its start on a farm somewhere around the world.
There are 1.9 million farms across the country, and 95% of those are operated by individuals, family partnerships, or family corporations, according to the 2022 Census of Agriculture from the U.S. Department of Agriculture. Despite our obvious reliance on these farms, they’re facing a labor crisis.
In Kansas alone, a study from Kansas State University and the Kansas Department of Agriculture found that between 5,627 and 15,873 jobs in 72 agriculture and agriculture-related sectors are unfilled.
Farmers are not strangers to technology, but this labor shortage is pushing them to automate more of their operations. CNH has more than 180 years of experience working with farmers, and now provides automated equipment to make farms more efficient and improve farmers’ lives. The company creates equipment for all kinds of crops in all kinds of climates.
“What’s unique about CNH and [agriculture] is we are one of the few companies that really services almost every type of farming area in the world,” Marc Kermisch, the chief digital and information officer at CNH, told The Robot Report.
The company invited The Robot Report to its Raven Innovation Center in Sioux Falls, S.D., to see some of this farm equipment up close. Located in the heart of the Great Plains, this facility is a space where the team can test and show off its equipment to customers.
“As we build advanced robotics here, we are always thinking about ‘How do we cost down the technology so it’s affordable and it’s market-right for different countries?’” Kermisch said.
While more automation is sorely needed in agriculture, a number of challenges are still holding back adoption. For one, farmers operate on razor-thin margins, and are always trying to get more out of each crop cycle, noted Kermisch. When introducing new equipment to their operations, farmers are putting immense amounts of trust in robotics developers, as one lost harvest could put the entire business in jeopardy.
“The other challenge is that most parts of the world only grow crops once a year. So, think about testing an autonomous tractor,” Kermisch said. “For testing advanced tech, computer vision, and machine learning, I’ve got to wait for that crop to come in. Then, I’m going to go through cycles. So it can sometimes take us two growing seasons to fully validate, which means the cycle of change and technology for farmers is a little bit longer than you would see in other applications in other industries.”
“So, one way that we try to manage our validation testing cycle is we have 200 acres here. This is all test farm where we grow crops,” he said. “Even when we’re not growing crops, we run simulations in the field.”
At the Raven Innovation Center, CNH showed off three technologies to automate different parts of the crop cycle.
While it may seem like the typical North American crop cycle starts in the spring, when farmers typically have a two to three-week time frame to plant up to 10,000 acres of crops, but it really starts much earlier with extensive planning for the growing season.
“In wintertime, [the farmers] are taking all the data that we can produce off of our tractors, they partner with their agronomist,” said Kermisch. “They’re saying, ‘What did I plant last year? What was the moisture that was produced by my farm or by the environment to grow my crop? What were the bugs? What were the weeds that impacted my growing and my yield? When I was at harvest, what was my yield?’”
Variations in planting can include seed spacing and depth, what fertilizers are needed for each plant, and the moisture levels of the soil. CNH offers Soil Command, which allows farmers to determine how much they’re going to turn over the soil and at what depth, broken down by a tenth of an inch, automating the process to prepare the field and better manage for variations in the soil across the field.
In the winter, after harvest is over, farmers also typically till their fields to prepare for next year’s crops. Turning over the soil helps to loosen the dirt, making it easier to plant new seeds. For this process, CNH showed me its autonomous tillage system.
The first thing that struck me about this system was how easy the interface was, which was true for all of the CNH technology I saw. With just a few clicks, the CNH team could input the parameters of the field to be tilled and which direction the equipment should move.
Once the tillage specifications are inputted, the autonomous tiller takes off with a honk, telling anyone in the field that the equipment will begin moving. Even though there’s typically no one in the field while the tiller is running, the system is still always on the lookout for obstacles in its path. If a person approaches the system, it slows down and then stops altogether if someone gets too close.
While the tiller is a massive and heavy-duty piece of equipment, it’s also incredibly sensitive. Part of reducing costs for farmers also means reducing maintenance costs, which go down when the system breaks down less often. To ensure this, the system can detect when it hits something in the ground, like a rock, that is too big for it to deal with. When it detects something like this, it stops and alerts the user to the issue.
Kermisch said CNH’s tilling product is set for limited release next year and a full release in 2026. Right now, the company is testing it with its customers and ensuring that the technology is ready for commercial release.
“So, we’ll be somewhere between five and 10,000 hours of tillage work on customer land this year,” Kermisch said.
“We have a new technology we call Augmenta that starts to identify the health of the plants,” said Kermisch. “It can help determine what and where and how much you need to spray.”
The technology can help ensure that farmers are giving each individual plant only the amount of fertilizer it needs. During the beginning of the season, when the crops haven’t sprouted yet, Kermisch said Augmenta searches the soil around the plant for weeds. It’s essentially searching for anything green against brown dirt.
While the system isn’t yet fully functional, he said it will eventually be able to continue identifying weeds as the crop grows in a process called green-on-green identification.
On the field, CNH’s precision-spraying technology uses Augmenta to care for each crop. The sprayer can run at 20 mph (32 kph) and can have booms up to 120 ft. (36.5 m) wide. The booms are outfitted with cameras, sensors, and adjustable spraying nozzles, giving the sprayer the ability to see the ground, take in information, and decide how much to spray.
“We have a vision base so we can identify the row and ensure we’re spraying within the row, and we’re limiting the effects of the spray,” Kermisich said. “We also have a variable rate nozzle control to determine how much is going out.”
The weed-identification process varies greatly depending on where the farm is located. In North America, there are around a dozen variants of weed types. In Brazil, however, there are almost 300, said Kermisch.
“So these algorithms really have to be purpose-built based off the region,” he said.
Harvesting is a crucial time for farmers. As with seeding, farmers have a very short time period to harvest their crops, and judging the best time can be a challenge on its own.
According to Kermisch, many Midwestern farmers need to wait until their crops are as dry as possible so they don’t rot while they’re being processed.
“You’re letting the sun dry out your plants, which then makes it more efficient to process,” he said. “This is especially true for corn, wheat, soybean, and obviously different for fruit.”
The challenges only continue once the actual harvesting process begins. Row crops, like corn and wheat, are harvested using a combine that cuts the crop. It’s important for the combine to keep running all day to harvest as much as possible, which means the combine needs somewhere else to put the crops it harvests.
To keep the combine running, farmers typically drive a cart next to it, which catches the processed produce as the combine processes it. Without automation, this process involves two workers, one in the combine and one in the cart, communicating with radios as they try their best to keep the equipment lined up perfectly.
If the cart is going too fast or too slow, it won’t catch all of the produce, and if the cart gets off track, and runs into the combine, farmers could be facing expensive repairs and a full pause in operations.
CNH simplifies this tricky process by automating the cart and letting the technology ensure the equipment stays in the right position at all times.
[See image gallery at www.therobotreport.com]The post How robotics can optimize every step of the growing cycle appeared first on The Robot Report.