News 11’s April Hettinger digs into the root of robots in weeding
YUMA, Ariz. (KYMA, KECY) – Technology is all around us. We have daily access to it right at our fingertips, and soon enough it is going to take over many tedious tasks in agriculture, like weeding.
Artificial intelligence is the way of the future, only getting better by the minute.
“This is a an AI based machine learning platform that has a very good ability to take pictures of plants or the crops we want to keep and open up mechanical blades around the plants we want to keep in close after the plant and cultivate and remove weeds,” said Ben Palone, senior technical product and product manager at Farmwise.
High-tech cameras are the eyes of the robot, and the brain is pretty close to what a farmer would be tasked to do.
“This is all based on digital camera technology. We take the pictures and then the pictures are fed into a computer, and the computer analyzes the pictures and decides what to do on the field after,” said Tony Koselka, co-founder of Vision Robotics. “So, going 1, 2, 3, 4 miles an hour, it takes a picture and quarter second later decides it’s turning on an office sprayer, it’s deciding whether to open a blade to weed around a plant or not, and is working that fast.”
The robots even know where the field starts and ends.
“They use a septentrio GPS unit that can take the coordinates down for the beginning and end of each bed,” according to Hunter Torres, regional manager of NAIO Technologies. “So, it follows a straight line down the bed and at the end, it knows which way it’s going to go. We set it up through the computer on the mapping software.”
So why does weeding matter?
“Weeding is important because one weed can produce 10s of 1000s of weeds from the seeds,” said Anthony Oddo, spray division manager at Grimmway Farms. “So, you got to get into these fields before they set flower and set seed, and with labor concerns, with the cost of labor and the quality of labor as well, these machines that you’re gonna see out here today are crucial for the success of farming.”
And, with an active drought, water resources could not be more important.
“Removing weeds: it’s obviously detrimental to the plants,” Palone explained. “It’s stealing nutrients and stealing most importantly water where water is a critical resource currently.”
Weeds give growers a year-round challenge, and with labor shortages, there aren’t enough workers to consistently hand-weed every field. That’s where technology comes into play.
“It’s really helped with consistency thinning compared to hand crews, and it’s also saved money on the farm on a cost-per-acre basis,” Oddo stated.
Each machine is customizable for the spacing of the beds and type of crop, according to Kim Haug, managing director of K.U.L.T. Cultivation Solutions.
“Everything’s adjustable on on each individual heights, you know, working depths, closeness. You can adjust all of it in the menu as far as how close to the plan you will go,” Haug said.
This includes automation machinery, too.
“Mapping the field, usually depending on the size, takes anywhere from 20 to 30 to 40 minutes,” Torres stated. “That just revolves around taking a point on each bed so the machine knows where it’s going, and it knows where to do the u-turns. Then we’ll come back into the next bed.”
One weeding robot was invented by two men in France, and now these popular machines split the year moving through fields in Salinas and Yuma.”
“These machines cost anywhere from $50,000 to $70,000 if you purchase them in elsewhere countries. Here in the u.s., We’re not able to sell them so we use them as a service here,” Torres said. “So as a service, it’s typically its cost per acre, and these machines are fully electric. They can do between six to eight acres in a day per machine for about a 10-hour run time.”
They plug into the wall to charge just like an electric car.
“These machines are fully electric driven. So, they’re very quiet so they’re easy to use. Especially here in yuma, there’s a lot of fields around residential neighborhoods,” according to Torres. “You don’t have to deal with the loudness of diesel tractor anymore.”
Other robots run entirely on their own, with one farmworker there for quality control.
“The machine is mostly autonomous with one operator and a tablet, walking behind the machine and keeping an eye on the on the quality perspective as the machine runs,” Palone stated.
Technology in ag has been in a trial and error process for around 50 years, but now they have just as much brain power as a computer, if not more.
“The type of algorithms and software that we’re running have previously only been used in software applications, and this is really kind of the first ability to use it in the ground for farmers,” Palone explained.
Preventative weeding has also shown to be effective with another machine that uses steam to run through the fields before planting, just like how University of Arizona extension associate professor Dr. Mark Siemens does.
“The purpose behind this is to heat the soil to a fairly high temperature about 140 degrees to kill the weed seed in the pathogen. As you might imagine, this is a pretty intensive process,” Dr. Siemens stated. “So, the idea we’re researching is the concept of band steam.”
The steam treats the seed line where it’s most important to kill the weed roots. The crops can be planted 24 hours after the band steam.
“You have some plants that we treated about four weeks ago and you can see some pretty visual differences between treated and untreated,” Dr. Siemens said.
Technology has come a long way, but robots are still only in their infant stages as a farmworker replacement.
“These are more consistent, and I think the technology is there, but in the next couple years, it’s going to be getting lighter and more efficient and you’re going to be seeing a lot more of this out on the farm,” Oddo explained.
Through a collaboration, farmers and computer engineers are on the right track to overcoming the never ending weed control and labor shortages.
