Smoke Some Weeds: Lasers Could Make Herbicide Obsolete

We’ve all tangled with unwelcome plant life at one point or another. Whether crabgrass infested your lawn, or you were put on weeding duty in your grandfather’s rose patch, you’ll know they’re a pain to remove, and a pain to prevent. For farmers, just imagine the same problem, but scaled up to cover thousands of acres.

Dealing with weeds typically involves harsh chemicals or excessive manual labor. Lasers could prove to be a new tool in the fight against this scourge, however, as covered by the BBC.

The Status Quo

Large liquid chemical sprayers like the Goldacres Crop Cruiser are the standard way many farmers deal with weeds today. Credit: Goldacres

Weeds cause all kind of problems for farming operations. They rob precious moisture and nutrients from the soil, and can crowd out the crops themselves from growing properly. In dire cases, they can contaminate food crops, making them unsafe for human consumption. To deal with weeds, farmers typically use giant spraying apparatus to hose their fields down with herbicides at regular intervals, both prior to planting and during the crop growth phase. These can range from small back-pack mounted apparatus, to gigantic self-propelled machines with 50-meter-wide booms that can spray 10,000 liters of anti-weed agents in a single day.

Recent developments in the weed-killing space have involved using camera systems to differentiate crops from weeds in the field. A variety of techniques are used, from imaging with multi-spectral cameras, to the use of AI systems. Conventional spraying machines can then be fitted with special controllable nozzles that target herbicide application onto weeds directly. This saves huge amounts of chemical versus spraying entire fields indiscriminately.

A Green Day

That same technology could help eliminate the need for herbicides entirely. Weeds identified by a camera system can be targeted with high-powered lasers instead of using chemicals to destroy them. One such system has been developed by Carbon Robotics, and is known as the LaserWeeder. Towed behind a tractor, the LaserWeeder is capable of killing up to 200,000 weeds per hour using an array of thirty 150W carbon dioxide lasers, operating at 10.6 um in the infrared range. Twelve high-resolution cameras are used to image the soil surface, identifying weeds for destruction. The system requires 60 kW of power to run, which is sourced from a generator run by the tractor’s PTO shaft. The apparatus is able to cover a 20-foot width, and is adjustable to suit a variety of crop row spacings.

It’s a simple idea that nonetheless requires great technological sophistication to pull off. The laser system must be accurate to a sub-millimeter level to hit the weeds in the right spot on the meristem to ensure a successful kill. The results are worthwhile, though. Laser weeding eliminates the need to use expensive chemicals on an entire field. Instead, the budding weeds are instantly destroyed by the laser, with their remains simply breaking down in the soil. It also eliminates the problem of spray drift, where the wind carries herbicide away from a targeted crop. As a bonus, the LaserWeeder can also thin out crops that have been overseeded, helping improve yields come harvest time.

Laser weeding is a no-contact method that could drastically slash chemical use in farming. Credit: Carbon Robotics

Laser weeding, along with all targeted weeding solutions, could come up against new hurdles in future, though. Evolution may yet rear its ugly head. Historically, some weed species have evolved a resistance to various common herbicides, making them far more difficult to deal with. The same could happen here. AI image recognition systems rely on unique visual features to determine if a plant is a desired crop or a hated weed. Thus, such a system could create evolutionary pressure on weeds in the field. That could help breed harmful weeds that begin to visually emulate surrounding crops for their own survival.

The technology does have one major drawback: speed. At just 20 feet wide, and running at a speed of 1 mile per hour, the LaserWeeder covers just 2 acres per hour. In comparison, rates of up to 200 acres per hour are possible with large self-propelled sprayers in the field. It bears noting that figure can be cut in half when accounting for the time taken to refill the chemical tank, but it’s still a huge point in favor of traditional spraying. In time, the company hopes to launch an autonomous LaserWeeder that will negate this problem, but it’s a ways in the future yet.

Currently, the company has focused on sales in the US and Canada, and has tested the LaserWeeder hardware on a variety of crops including potatoes, onions, and garlic. The AI image recognition system is set up to work with over 40 crops in total. Expansion into European markets is on the company’s agenda moving forward.

For now, laser weeding is a nascent field that could yet change agriculture across the world. Few farmers relish the days spent mixing up chemicals and babysitting their sprayers in the field, nor do they enjoy the cost outlay. A non-contact weed control method with less consumables would be welcomed as a game changer by many. Expect to see the practice of laser weeding explode in coming years if it proves as successful as its creators hope.

53 thoughts on “Smoke Some Weeds: Lasers Could Make Herbicide Obsolete

    1. Not only those two. Take DDT for example: it was forbidden in 1972 in US, in 78 in EU. Try to find a breastfeeding mother that not not have some in her milk, 45 years after. And also organochlorides, PFAs, dioxins (…)

      Those millions of metric tons of poisons spread on our culture diffuse in the whole environment, following the water cycle, and concentrate on the top of the food chain. And who is on the top of the top of the food chain ? human babies. Human milk is poison. You doubt it ? Test it. And prepare for a shock.

      If they don´t kill, they cause cancer, hormonal imbalance, sterility (…)

      Insects. Those who are over 40yo remember windshields splattered of dead insects. And now ? almost noshing. Less insects, less birds, less plants less biodiversity, more unbalanced ecosystems and ultimately: collapse. Those who survived became resistant, thus more pesticides are needed for the same effect.

      This is the new normal and nobody cares.

      And the BS argument that we cannot feed the planet if not using pest- herb- fung- icides is … BS. True it´s not possible for extensive agro-production without those poisons. It´s also not possible to produce for that cheap without those poisons. That cheap price is masking the ecological price, and this will be really, really heavy on the future generations.

      So, this laser gadget might help using less poisons, but ultimately it´s the whole industrial food production that we must change. Not the tools or the chemicals, but the way of producing.

      1. @ono: “Those who are over 40yo remember windshields splattered of dead insects. And now ? almost noshing. Less insects, less birds, less plants less biodiversity, more unbalanced ecosystems and ultimately: collapse.”

        That !

  1. 2 acres per hour, while requiring 60 kW of electric power. That’s going to be around 12 litres per acre of diesel fuel. AND it ties up a tractor for the duration, to haul it around, doubling the fuel burn.

    Add amortization/rental and labour costs, you are looking at a minimum of $200/acre cost, which is about the same as the gross margin on common crops like soybeans and corn.

    This thing has a tough row to hoe.

    1. 60kw is 80Hp plus generator losses. It doesn’t take another 80 hp free m the tractor to pull a rolling 20’ wide implement at 1mph, so fuel cost is less than doubled.
      The cost of herbicide and application is removed, 25-100$/acre, so your estimate of $200 vs gross profit needs some tweaking. If there is a premium price at market for no herbicide crops, that would improve the profit margin as well.
      But thanks for bringing up the profitability aspect.
      An autonomous electric version that drives at night and charges from solar during the day will be good. My farmer just put in 44kw of solar, like lots of his neighbor.

    2. Using lasers to reduce the population of bugs, especially flying ones? I like it. Using them to kill weeds in a crop field? Like you say, no way is that going to work out economically favorable, and if it does, I still don’t want to burn all that extra diesel. For all I know, you’ll have to go around with this rig more often than normal too. Seems like we’d be better off with human or robot arms, if we can’t agree on a tolerably safe herbicide or alternative practices to reduce the weed problem.

        1. There problem here is that this system is a joke. It is not competitive in any way save for removing herbicides and will not actually work as described in the first place. I am all for alternatives, but this is not one.

      1. Pollution-wise, you’re far better off burning a little extra diesel. The alternative is spraying thousands upon thousands of liters of weed killer… and you’re still burning diesel while you drive around doing that, too.

        1. There’s plenty of ways to apply weed killer / pesticides / fertilizers / etc – center pivots can spray them out mixed with the water, which is the lowest energy option I can think of at the moment. But even if you used a tractor, at least you don’t burn several times as much diesel by going incredibly slowly and driving a generator for lasers as well. I believe spray rigs are generally available in up to a very wide swathe and can be run at good speeds, so many more acres per hour. A challenge of course is picking something acceptable to spray.

    3. Couldn’t you just make a system of lenses and mirrors to use the sun as the energy source? Then you only need the tractor to propel it around the field (which you have to do with the weed sprayer, anyway). The targeting system – and maybe even the propulsion – could be photovoltaic-powered.

      1. I don’t think you could do the same thing, even with a really good system of mirrors and things. You can concentrate the sun to a fairly small spot with a perfectly aimed lens or mirror, but you can’t make as tight of a beam with it as a laser, because lasers tend to have much lower divergence than other sources.

        1. But you only need it focused at certain point say 1cm2 you dont need kilometers long beam. And that you can do with optics simple magnifying glass is enough. you’d need very fast clever sun tracker for moving vehicle and it doesn’t need big array of mirrors in climates with lots of sunny days, only for places like UK you’d need bigger array due to lots of cloudy days.

  2. I think indoor farming is going to have a bigger impact over all than this laser system. That being said there are some crops that don’t do well in hydroponics or aren’t economically viable for indoor growth; return on investment for a 500 acre green house will be long. If they can boost the efficiency of the lasers, and work out the kinks in the autonomous platform this idea has real merit.

    Another approach they could take would to adapt an irrigation pivot with a gantry that the laser system can travel on. The power plant would be stationary at the pivot point, which is less weight to haul around, and path finding / navigation would no longer be an issue. Given time they might be able to keep the laser generating system stationary at the pivot as well, using fiber-optic or shrouded beam paths to transmit the energy out to the scanning assembly.

    1. With gantry you could be seeding at location and don’t use laser to kill but just servo with spike to kill weeds and with location of seeds you don’t need advanced AI to determine if its crops or weed it its not at X/Y location its a weed smack it.

      1. So… Farmbot ( ) only on a massive scale. We’ll lease the AI visual weed ID system for the cool factor to draw investors, then add in the spike/claw bit because lasers aren’t cheap. Farmbot is open source which solves the software problem. We’ll just need to convert it to run on an radial pivot instead of a parallel tracks.

        Know anyone good at running kickstarter campaigns? We’re going to need some seed money for graphic artists and to build a miniature system on a one acre field. I’ll be in Austin next week, those nerds have more money than sense, we’ll be on the cover of Forbes by this time next year. That, or broke with a lot miniature robotic farm implements.

  3. Just killing the apical meristem on a weed? That’s not going to kill the whole weed. Those things are terribly resilient — if the apical is killed it very likely has many many lateral meristems from which it will proceed to live long and prosper. My experience years ago walking beans, the only way to go with most weeds is to chop them off at the soil surface and even then some will regrow.

    1. What about just apply high voltage electricity to the weed? I have done that with a flyback transformer, it works but takes a while to destroy the root due to the low current, something like 5mA. Increase it a few times and it should be much faster, while power required is still in the few tens of watts.

      I don’t know the exact mechanism by which it destroys the root, I would expect the current density to be too low to do that by thermal means. My guess is something to do with the current causing a chemical reaction.

  4. It looks like a good idea until you put Economics on the equation. It’s gonna be so expensive that makes unprofitable to run a farm on those lasers.

    If you really want to use power and light to farm, expend that power on LED lighting for urban farming (that are not that profitable right now).

    Or use power for several greenhouses and use hydroponics, no need to kill weed when there’s no weed to start.

  5. I wonder how good the cameras would be at targeting drones, given the right training data. Turn this thing upside down, multiple lasers working together.
    Maybe train it to take out mosquitoes.

  6. In Australia we have days where the use of all outdoor fires, and potential ignition sources, are banned, I suspect that his tech would have that rule applied to it.

      1. Doesn’t matter where you want to apply it as the government would make it unlawful to deploy anywhere on certain days and that is totally out of your control so hard to manage. Not impossible, but it does add to the complexity of managing the production sequencing in agriculture.

    1. Agreed. Considering some of the 2019-2020 fires were added to by spot fires starting from the hot exhaust of a farmers ute, this contraption would be a no go in NSW during total fire bans.

  7. Could this be scaled down for household use? A backpack (or roll around two wheel cart) with a fiber optic wand that plugs into a wall outlet?
    I have a bindweed (with roots that go down to Hell) problem from Satan himself.
    Hand picking? Oh, my aching back!
    Sprays? Toxic.
    If this could burn the roots…….

    1. A battery in a backpack and a small laser-engraver head on a stick with a switch seems like the quick & dirty version of this – no idea how much power is needed or if there’s a specific wavelength.

      Then again, one of those weed burner wands that just takes a small blowlamp gas cylinder does the same thing and is on the shelf in your local hardware store for very cheap.

        1. I stopped caring about dandelions when I learned that every part of the plant is edible. Add the fact that they don’t have thorns or spines (like some of the other lawn “weeds”) and now I can tell myself that my lawn has just become more “useful.” But yeah, steam would still be handy for thistles and bindii…

  8. Everyone thinks it’s SkyNet which starts it. But THIS is how it all was started. First it was just lasering weeds to elimination, it was takes too long so AI and automation was adapted and soon after the AI figured out it was the humans who were the ‘weeds’ on planet Earth. That’s when they start zapping humans.

    1. That reminds me of the Lost City of the Exxilons in the Doctor Who story “Death to the Daleks”. The Exxilons were once a highly civilised people, who decided to build the perfect city. It was to be capable of sustaining itself, finding its own power and resources, and even improving its design. When it was turned on, it realised that it had a flaw: it was full of people. It turned them out and built a wall around itself. When the story starts the world around it is a wasteland, with all the good things sucked up by the city’s “roots”. It also absorbs any concentrations of artificial power, so the citizens have been knocked back to the Stone Age. But if you can get into the City, past its defences, it’s very nice inside.

  9. Maybe recognize the crop and kill all the rest. Crop is in a row weeds may be at random. A little spinning toothed wheel poked at the bottom of weeds will consume far less power than a laser. Things don’t have to be nearly as accurate to kill. Basically a visually smart articulated cultivator, that makes sense. Lasers. Really. There is also propane flame weeding and bug killing going on out there. Something else comes to mind reading about potato bugs getting shaken off or vacuumed off and getting zapped with HV.

  10. Solar panels
    Autonomous vehicle
    AI targeting
    Green house agriculture
    all parts of the solution

    the more people who do this the less sources of weeds there will be
    lets stop poisoning ourselves with herbicides that don’t just kill plants

  11. This article totally made me think of one of my favourite books, The Turing Option by Harry Harrison and Marvin Minsky, to tell more or why would be a spoiler but it’s worth a read.

  12. Lasers don’t really need a tractor either. Could do a center pivot with a carridge that travels up and down the beam so you don’t need so many lasers and cameras. Hook it up to a solar array/other renewable source of power and just leave it plodding away.

    Of course the problem of the initial costs being considerably higher than a spray sytem still remain.

  13. I am not sure that the complexity and expense of a laser is needed. A good shot of electricity could do the job too. The weed does not need to be fried, a modest jolt would probably be enough.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.