Agriculture
Better than Pesticides: Laserweeding
by Laura Fitzgibbons
In a field of leafy spinach and carrots, tiny dots of light sparkle across the soft dirt. They are not fireflies
in the warm New Mexico air, they are laser beams hitting the soil at lightning speed.
Above the flashing lights, a giant, self-driving tractor glides at a snail’s pace of one mile per hour down a
perfectly straight path. Futuristic farms like this, where technology works in harmony with nature, are
on the rise and may be the answer to many of the challenges facing agriculture.
Pesticides have been used in agriculture to keep weed populations in check for centuries, despite their
many drawbacks. They may stop crop plants from properly growing, create abnormalities, stop cell
division, and even kill the plants. The harm from herbicides extends beyond the plants and into the
water supply, soil, and can hurt humans who eat the crops. The chronic effects of pesticides vary, but include
birth defects, immunotoxicity, and many forms of cancer. Researchers and farmers have long searched
for alternatives to this practice.
Additionally, labor shortages in recent years have created unprecedented struggles for the agriculture
industry. Retention and recruitment of agricultural employees is worsening.
As Manoj Karkee, a professor at Washington State University and agricultural engineer explains, “We
still use a lot of human labor, and we don’t have enough people available to work in these farm
operations anymore."
He goes on to explain that in the US, “it is difficult to find people to work in the farm, and on top of
that, the conditions for laborers are often tough, from heat and long days in summer, to cold operations
in winter, and even the smoky conditions from wildfires." There are many challenges to farm work, but
according to Karkee, “with advancements in AI and robotics, I think we deserve better, and I think
we have capabilities now to develop technologies that perform risky, dangerous, difficult jobs while
providing opportunities in other areas such as manufacturing, supervising, operating, and maintenance."
Labor shortages come amid an increased demand for food globally. According to the World Resources
Institute, Earth’s population will be 10 billion by the year 2050. This increase could mean 593 million
more hectares of land dedicated to agriculture, and at least 56% more food production to sustain that
population.
The efficiency of real-time farming decisions with AI has several benefits. Artificial Intelligence
technology helps to make effective decisions like how much water a specific type of plant prefers,
where to place a certain crop on the farm, and detecting diseases and insects.
According to Manoj Karkee, "Agriculture is an industry in my opinion that is so critically important for all
of us because we certainly have to rely on the fuel, fiber, and food that is produced through different
kinds of agricultural industries. We’re talking a lot about how to safely use these advancements for
overall societal benefit in a positive way...that kind of becomes a challenge for any new technologies
that we see invented in this world, particularly so with AI because it has such a widespread and really
significant impact it could make in the way we live, the way we work, and certainly it does this in every
walk of life.”
Manoj Karkee's research includes creating small robots that mimic the pollination habits of bees,
working to improve the health and quality of crops, and doing so in a non-invasive way that operates in
harmony with the actual bees in the ecosystem. These small robots can tackle big problems but keep
bees in business as they battle the harm of pesticide use including colony collapse disorder, and for
some species, even critical endangerment.
James Johnson is a farmer whose family has grown chili peppers, onions, and cotton for 104 years across
four generations. In a large sunny warehouse wearing a ballcap, he describes how his approach to
farming has changed over the past several years.
When he started farming, he was what he describes as a"conventional" farmer. In the beginning,
he only wanted to transition to organic farming, but he found that it is difficult to go fully organic. His
biggest hurdle in the beginning was keeping up with weeding using only organic materials. According to
Johnson, “any chemical has negative effects on plant or soil health,” and he needed a solution to
remove weeds quickly without herbicides.
Manoj Karkee sees agriculture tech heading in a positive direction for several reasons. "Now with AI and
modular robotics technology, I think we can provide tools such as cell-phone-based applications, small
robots, even drones.” This focused level of technology that is applicable to any location, scale, and
applications, has only recently become possible in the past few years.
One of the additional hopes of AI for agriculture is that it will have the ability to remember and store
oral histories about the day-to-day work of farming. According to Manoj Karkee, “Current farmers and
leaders in the farming industry have a lot of experience and knowledge around how farming is
done...and it’s important to preserve the knowledge and experience of the current farmers for future
generations.” AI may serve as a tool to preserve that expertise for future generations of farmers.
“In general I just would like to mention that these technological advancements have been and will
continue to be a powerful force in producing high quality nutritious food,” he says, “to continue to feed
the increasing global population, but even more importantly, these technologies will allow us to be
capable of doing so more substantively so that we can address the needs of our generation, but also
care for the environment and long term sustainability of these agricultural operations so that our future
generations have just as much of an ability to feed themselves.”
A new approach to spraying pesticides on crops may also drastically lower their use. A company called
Precision AI has created a system that uses AI to focus aerial sprayers mounted on drones on areas
needing weeding, leaving other areas untouched. The ability for the AI software to recognize weeds
within fields of crops with spot precision and remove them is called “green-on-green"; spraying. The
technology lowers water waste by up to 95% and eliminates soil compaction—the crushing of soil that
removes helpful microbes and oxygen.
Precision AI's drones use cameras and onboard processing to recognize weedy areas and adjust
herbicide application, passing over or lightly impacting areas needing less herbicide. While farmers still
need to apply herbicides in this process, they can use significantly less because of this advanced tech.
Precision AI’s tech applies autonomous optimization of aerial pesticide spraying right from the sky. Using
real-time data processing, this saves money and leads to better quality crops and less chemicals.
According to Amr Omar, the Head of Artificial Intelligence from Precision AI, it is important to “bet on
the process and not on the product while working in machine learning or in AI-driven teams. The
process is way more important than the product as the product will come at the end of the day.”
For their researchers, the process meant finding ways to improve weeding and tackling the physical
challenges of hard work on a farm coupled with the delicate hardware within a drone. This meant being
able to plan for things like clouds, rain, dust, and many other factors that can disrupt cameras, imaging
software, and sensors.
Another new technology has found a way to eliminate the use of pesticides. The Laserweeder by Carbon
Robotics uses a precise system—a combination of cameras, artificial intelligence, and high-powered
lasers—to designate weeds for removal. Once the weeds are spotted by the on-board, hi-res cameras,
they are destroyed with the precision-focused laser down to the roots, and in some cases, before they
even start to seed.
When farmer James Johnson tried the Laswerweeder, he could fully weed his fields with minimal
damage to the crops or soil.
The self-driving weeding tractor can run night or day, and despite its slow pace, can manage 100,000
weeds per hour or more. The system is the first of its kind, according to Carbon Robotics CEO Paul
Mikesell.
Invasive weeding and herbicides degrade soil health, while the Laserweeder leaves soil intact. This
improvement is evident on a microbial level, as Mikesell explains, "We don't tear up the dirt, so it leaves
that microbial bacteria in there that's there to fix nutrients and nitrogen. We dont spin that up, so we
let that ecosystem go and keep going, so that's good for your crops and it's much better for healthier
plants."
In spring of 2023 the company also expanded its AI capabilities to include Laserthinning, the practice of
remove extra plants from intentionally overseeded land. The technique is comparable to the tractor’s
weeding capabilities, but instead of locating and destroying weeds, the tractor can be set to find areas
with overcrowded plants and remove a portion of those crops, leaving enough space, water, and
nutrients for the remaining plants to thrive.
The Laswerweeder’s AI was trained with deep learning and repetition, observing pictures of crops and
common weeds until over time it learned how to differentiate between the two categories. The system
uses powerful carbon dioxide lasers that, while compact, have 150-watt strength, the equivalent of
lasers that can cut wood and metal.
Advances are happening throughout agricultural tech at the intersections of research and development,
computer science, AI, agricultural engineering, crop scientists, and socioeconomic scientists. Manoj
Karkee believes that there is strong potential for positive impacts in terms of producing food
geographically close to where it will be eaten, creating small-scale controlled environments for growing
food, vertical farms, rooftop farms, micro and city farming that might help preserve crop yields, and
hopefully these advancements can minimize hunger for everybody around the world.
Tackling massive problems like how we will survive as a species and a planet in the next several decades
is no small feat, but this tech paves a way for safer food, better land use, and less destructive weeding
practices.
An enormous tractor with no human driver, lumbering day and night across the country
fields—one that will instantly and precisely blast any weed in its path—may seem disconcerting, but
unless you are a chickweed patch in an onion field, you probably have nothing to worry about.