Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Robotics AI Businesses Earth

AI-Enhanced Weed-Killing Robots Frighten Pesticide Industry (reuters.com) 176

Rick Schumann writes: A Swiss company called ecoRobotix is betting the agricultural industry will be willing to welcome their solar-powered weed-killing autonomous robot, in an effort to reduce the use of herbicides by up to a factor of 20 and perhaps even eliminate the need for herbicide-resistant GMO crops entirely.

The 'see-and-spray' robot goes from plant to plant, visually differentiating the actual crops and weeds, and squirting the weeds selectively and precisely with weed killer, as opposed to the current technique of using large quantities of weed killer like Monsantos' Roundup to spray entire crops.

Weeds are already becoming resistant to such glyphosate-based herbicides after "more than 20 years of near-ubiquitous use," reports Reuters. (The head of one pesticide company's science division concedes that "That was probably a once-in-a-lifetime product.") But AI-based precision spraying "could mean established herbicides whose effect has worn off on some weeds could be used successfully in more potent, targeted doses."

Meanwhile, another Silicon Valley startup has built a machine using on-board cameras to distinguish weeds from crops -- and was recently acquired by the John Deere tractor company. Reuters calls these companies the "new breed of AI weeders that investors say could disrupt the $100 billion pesticides and seeds industry."

The original submission asks: Should we welcome our weed-killing robotic overlords?
This discussion has been archived. No new comments can be posted.

AI-Enhanced Weed-Killing Robots Frighten Pesticide Industry

Comments Filter:
  • Why spray them? (Score:5, Informative)

    by Snotnose ( 212196 ) on Saturday May 26, 2018 @11:40AM (#56678620)
    Seems like it wouldn't take much more mechanical engineering know how to pull them out by the roots. I have yet to hear of a weed that can resist being pulled out of the ground and tossed into the compost bin.

    Some weeds break at the soil line and the roots regrow, you say? It's a robot, run it again in a month or two. Extra added bonus, this time you get all those young whippersnapper weeds that were but tis a seed last time the robot came by.
    • Re: (Score:2, Interesting)

      by Anonymous Coward

      Add a water tank, heat the water with the sun, spray them with hot water.

      It's much easier mechanically.

      • by HiThere ( 15173 )

        That's an idea that should be considered seriously. It probably will be, eventually, and it sounds a lot more practical that pulling them. Pulling weeds takes considerable force, insulating a container of hot water a lot less. It also takes less delicate manipulators. Etc.

        But using weed killer first is reasonable. It could get easier market penetration. The weed steamer (boiler?) could be sold later to the organic farms at a premium price.

        • by djinn6 ( 1868030 )
          Force isn't a problem. Energy is. It takes much more energy to heat up a cup of water than lifting the weed a few inches off the ground.
        • Or an electric needle, quick shot - and although some weeds recover from that, as the op said, just send the robot around sometime later.
      • Anything solar powered is going to have trouble outside the tropics and near-tropics. That 1kW per square metre that people love to quote is only valid when the sun is directly overhead at midday and outside of that atmospheric attentuation means that it drops off with the cosine of the sun's angle to the ground (effectively this means that you only get viable power in the 2-3 hours each side of local solar noon)

        Unless it has a stupidly large solar panel, a robot trundling down fields is going to spend a l

    • Re:Why spray them? (Score:5, Informative)

      by Locke2005 ( 849178 ) on Saturday May 26, 2018 @11:52AM (#56678662)
      Many weeds like thistle break off and grow back from the roots. (Thistle amazingly has 3 foot long roots for a plant only a few inches above ground.)
      • Many weeds like thistle break off and grow back from the roots. (Thistle amazingly has 3 foot long roots for a plant only a few inches above ground.)

        As he said, run it again. It's a robot. What's it going to do, strike?

      • by tlhIngan ( 30335 )

        Many weeds like thistle break off and grow back from the roots. (Thistle amazingly has 3 foot long roots for a plant only a few inches above ground.)

        The thing is, weeding is extremely hard. If you can get a robot to do it even on a weekly basis, even weeds that evolved anti-pulling measures will have hard pressed to survive. It takes a lot of resources to pop a plant above the surface, and if it keeps getting chopped off every few days, eventually even the long roots will run out of stored energy (food) and

    • have a headline like "see the new gadget that terrifies the pesticide industry", mention magnets maybe and the investment opportunity. click bait on slashdot

      • If you want to be truly terrified, just wait till the weeds develop AI, and then watch!
    • Comment removed based on user account deletion
      • by gtall ( 79522 )

        Or Japanese honeysuckle, that stuff will strangle you given half a chance.

      • by Jeremi ( 14640 )

        You've never met Kudzu. Go ahead. Pull it up and come back a month later. The vine may be 30 feet long by the time you get back. It grows so fast, it's dangerous to sleep with your windows open.

        If the robot needs to come back and pull the weed again the next day, it can do so. If it needs to come back every single day for the entire season to pull it again, it can do so. Robot's got nothing but time.

    • by gweihir ( 88907 )

      It is a stage of development. Pulling it out is far more difficult than spraying it. You cannot have the final solution in the first viable product.

    • Last year, I visited a small vineyard. They sort of did that: they had a robotic lawnmower on constant patrol.
      Of course that only works if your crop is tall enough not to be harmed by the lawnmower.

    • by jsrjsr ( 658966 )
      One of the engineers I worked with told about working on an experimental electrical weeder for cotton. A charged bar was mounted on the front of a tractor and driven through the field. The bar was adjusted to be just a little above the cotton. When weeds grow, they typically stick up above the cotton, hit the bar and are electrocuted. He said it worked great, except for the safety problem. At the farm where they were testing, the farmer's dog came running up and hit the bar. So much for testing at tha
      • by jtgd ( 807477 )
        Couldn't they use a timer so it only powered on for an hour in the middle of the night? I would think a couple seconds would be enough.
    • Some weeds break at the soil line and the roots regrow, you say? It's a robot, run it again in a month or two.

      Yeah, talk to experienced gardeners about that. You have to either pull up the root, or keep cutting the new growth every week. A month is too long; as soon as the first new group of leaves start producing energy, the weeds put that energy into restoring the stored energy in the root necessary to put up another shoot. After two weeks, it hasn't just restored it's energy, it has also started to extend it's root system! If you wait a month, you don't just have another shoot to pull up; you went from one weed

      • it hasn't just restored it's energy [...] extend it's root system!

        It sounds like this stuff spreads like apostrophes in an imbecile's post.

               

    • Seems like it wouldn't take much more mechanical engineering know how to pull them out by the roots.

      A better strategy might be to use a tiny blade and cut the weed off just below the soil level. Some plants will regrow, but as long as they do it regularly the roots will eventually weaken and die. It’s basically the same thing as hoeing your garden regularly, which works great if you’re willing to stick to it (note to self: be more consistent about it in your own garden).

      I agree that it’s silly to keep spraying herbicide, especially since they’re visiting and identifying each plant

      • by ras ( 84108 )

        A better strategy might be to use a tiny blade and cut the weed off just below the soil level.

        This isn't exactly a new idea. Attach the blade to a stick, and it's called a hoe. Here is a picture of one used 2000 years ago [wikipedia.org]. The idea was used prior to that with wooden blades.

        A few 100 years ago we automated hoe by dragging a blade through the soil behind a cow or horse. The blade changed it's name to a "plough". In modern times we've changed to using a tractor instead of a horse.

        The technique of cutting a

    • Sure. But that's seventh century (BC) technology. Where's the pizzaz? The Sizzle?

      I think a tiny hoe might be easier to implement and more effective than uprooting. On top of which any patents on the process probably expired about the time Moses was leading the children of Israel out of captivity.

    • Not only that you could likely genetically engineer your plants to have some proper tracer that identifies them to the robot. Then you could do away with basically all that deep learning figure out what type of plant it is nonsense. Just make a robot that cuts down all plants without that marker. Just needs to be cheap to detect, etc. Then your computing power is down to what basically amounts to nothing at all and you can just cut them off at the soil line each time they emerge without said marker.

    • Also, if we could get these murder organisms robots working telling them to hunt down and kill all the bioinvaders of this or that variety would be nice. Murder all the first that look like this fish. Would be wildly useful for saving the environment. Or something like hunt down and murder every rabbit on this entire continent.

    • by Khyber ( 864651 )

      "I have yet to hear of a weed that can resist being pulled out of the ground and tossed into the compost bin. "

      Plenty of them growing in most compost bins that sit unturned for longer than half a week.

    • by pnutjam ( 523990 )
      I think you underestimate how fast weeds grow, after a month, weeds could be huge. Some weeds can also have roots up to 10 feet deep.
  • Resoundingly YES (Score:5, Insightful)

    by Kobun ( 668169 ) on Saturday May 26, 2018 @11:51AM (#56678658)
    This is an awesome job for AI robots. Either:

    1. It replaces hard-on-the-body manual labor (long hours bent over, sun & sunburn & skin cancer, etc) by using a robot to pull weeds.
    2. It delivers targeted doses of herbicides, hopefully reducing the enormous amount of Glyphosate(*) currently used AND reducing Monstanto's ridiculous amount of control over the farm industry.

    * - Over 90% of all glyphosate produced and used EVER has been in the past 20 years. 70% in just the last 10. Food today is NOT the same as it was in the 70s, 80s, and early 90s. Do you trust Monsanto to proactively limit the amount of Roundup you consume?

    https://www.ecowatch.com/monsa... [ecowatch.com]
    • Over 90% of all glyphosate produced and used EVER has been in the past 20 years. 70% in just the last 10.

      Duh. Couldn't very well spray it on crops until we developed roundup resistant crops.

      Do you trust Monsanto to proactively limit the amount of Roundup you consume?

      No, I trust my faucet.

    • by gweihir ( 88907 )

      I fully agree. This is a very good idea. Eventually, it may be completely poison-free, but obviously that is a future stage of development.

  • Fertilizer? (Score:4, Insightful)

    by cirby ( 2599 ) on Saturday May 26, 2018 @11:59AM (#56678698)

    Just apply very concentrated doses of fertilizer and other "good" soil chemicals - enough to poison the weeds when applied directly, but good for the crop when diluted by irrigation or rain.

    • Just apply very concentrated doses of fertilizer and other "good" soil chemicals - enough to poison the weeds when applied directly, but good for the crop when diluted by irrigation or rain.

      concentrated fertilizers will also kill your seedlings you're trying to grow

    • Re:Fertilizer? (Score:5, Insightful)

      by theCoder ( 23772 ) on Sunday May 27, 2018 @06:03AM (#56682238) Homepage Journal

      Where I live in Florida, fertilizer is considered a pollutant. This is because when it rains it collects in the waterways. There, it does its job by getting plant material to grow in the form of algae. These algae blooms can get so thick that they end up killing the animal life. Sometimes, the blooms are toxic to humans and create really bad smells.

      The point is that fertilizer is not always good and concentrations from run-off can have unintended consequences.

  • by Anonymous Coward
    I wonder when will weeds starts imitating the look of crops or maybe grew some patches/patterns that triggers the false positive of the image recognition algorithm and avoid being picked up by the robot... Sounds scary.
    • by careysub ( 976506 ) on Saturday May 26, 2018 @01:15PM (#56679022)

      This is called Vavilovian mimicry and is very, very well known. There are many weed species that have been selected to match the growth period, habit, and appearance of crop plants, and are thus propagated inadvertently by farming. They only need to mimic during the part of the their life cycle where they are subject to weeding, but some are near replicas are of the crop plant despite belonging to different genera.

      I think robotic farming may make this a lot harder for the weed. If you have robot planter that exactly space the seeds, and robotic sprayers that can both recognize crops based on their appearance, but can also use their planting pattern for recognition, it could become a lot harder. Also the optical sensors could out perform the human eye. Narrow band filters might be very useful for recognizing the crop plant. In fact this might offer another genetic engineering tool. Add an unusual pigment or pigments that reflect specific wavelengths which the robot can detect with filters, but won't be found in any of the weed species. Essentially adding optical tagging.

      • by ras ( 84108 )

        +1, excellent post.

        Mind was even more blow when I look it up on Wikipedia, and discovered Rye was once a weed that did Vavilovian of wheat, and imitated it so well it became a food crop itself. And that isn't the only food crop that came about like that. Amazing.

      • Narrow band filters might be very useful for recognizing the crop plant. In fact this might offer another genetic engineering tool. Add an unusual pigment or pigments that reflect specific wavelengths which the robot can detect with filters, but won't be found in any of the weed species. Essentially adding optical tagging.

        Not to ride on the coattails of your eventual Nobel or anything, but this is the best idea I've heard in a long time, and I spent five years at a crop-science company. Put together a robust trait stack to express three or four distinct narrowband fluorescent emissions, and good luck horizontal-transferring that entire stack into a neighboring weed by accident. If the robots see something that only fluoresces in one or two of the bands, it'll be as easy to pick out as a purple or cyan blot in a field of whit

    • by HiThere ( 15173 )

      Too late. There are weeds in India (at least) that have started imitating the look of rice plants to prevent people from pulling them out.

  • can we train it to weed out politicians ?

  • by hibiki_r ( 649814 ) on Saturday May 26, 2018 @12:33PM (#56678836)

    Robotics can make sense in agriculture, but as every farmer will tell you, farming is really about economics: Getting good returns of investment on a weeding robot, given how relatively cheap pesticides are, is going to be tricky given regular year's crop prices.

    The last time we had a boom in agricultural tech had nothing to do with the tech getting better, and a lot with a terrible drought in the midwest that made prices skyrocket. When corn pays $8 per bushel, instead of $3, suddenly everyone was willing to buy tech. Then prices drop again, nobody wants to keep buying, and many people that bought into the tech when they expected $8 forever just went bankrupt.

    Eventually we'll have cheap enough robots that will also handle pests, and might even isolate or destroy diseased plants, lowering our chemical uses while still having great yields, but in agriculture, it's the economics that is stopping a tech boom, as aiming for the top of the technology chain, including the best seeds, is quite the gamble given what weather can do to your operation.

    • by c6gunner ( 950153 ) on Saturday May 26, 2018 @12:58PM (#56678956) Homepage

      Ah, but that's the whole benefit; if you use robots then that $3 bushel of corn which used to sell for $8 during the price spike can now be sold for $12 to gullible rubes because you can stick the word "Organic" on it.

      Of course if everyone starts using robots instead of pesticides then the price difference drops and the farmers are no better off than they were when they started ... but, in the meantime, why not take advantage?

      • by HiThere ( 15173 )

        Doesn't work if you're having the robot spray weed killer. Your fields then don't qualify as "organic". (Steam or microwaves would work, though.)

      • by raymorris ( 2726007 ) on Saturday May 26, 2018 @01:54PM (#56679214) Journal

        That's funny. What's worse is that the "organic" label in no way means "no pesticides". It means they used what they call "organic pesticides", which are pesticides that are chemically similar to some of nature's more potent toxins. What's not allowed on organic produce are the newer, more targeted pesticides which designed to be effective against insects but harmless to people. Instead, organic pesticides are based on the chemistry of toxic plants.

        It's like using belladonna (deadly nightshade) to treat ulcers. It works if you happen to get a belladonna plant with just the right concentration of hyoscyamine, and you take just the right amount, but it's a hell of a lot safer and more effective to use modern compounds like amoxicillin or Prilosec.

        In one recent USDA study, lettuce marketed as organic contained, on average, ten times the amount of pesticide as lettuce not marked organic. That's because "regular" lettuce can use trace amounts of modern, much more effective, compounds, rather than drowning the lettuce in a toxin that's naturally produced by a bacterium.

        • That's funny. What's worse is that the "organic" label in no way means "no pesticides". It means they used what they call "organic pesticides", which are pesticides that are chemically similar to some of nature's more potent toxins. What's not allowed on organic produce are the newer, more targeted pesticides which designed to be effective against insects but harmless to people. Instead, organic pesticides are based on the chemistry of toxic plants.

          Yeah, toxic to bugs. Did you know that mint oil is a neurotoxin to wasps? You're ignoring half of the argument in order to demonize organics.

          • Yeah, toxic to bugs. Did you know that mint oil is a neurotoxin to wasps? You're ignoring half of the argument in order to demonize organics.

            No, he's not. Some stuff - like mint oil - is mostly just toxic to bugs. Other approved "organic" pesticides - like copper sulphate - are toxic to pretty much every type of animal life. Some approved pesticides, like pyrethrin, are known neurotoxins. While pyrethrin tends to break down relatively quickly some residue still remains on produce, and it certainly poses

            • No, he's not. Some stuff - like mint oil - is mostly just toxic to bugs. Other approved "organic" pesticides - like copper sulphate - are toxic to pretty much every type of animal life.

              Yes, he is, because he's conflating them all. Which ones are used in which percentages, and which ones are easiest to wash off are both relevant.

              When I've done organic gardening, I've used literally nothing which was harmful to humans. Castile soap, neem oil, and baking soda cover the majority of problems. Sometimes I had to bring in some insect predators, but they turned out fine. Not using pesticides which harm beneficial insects means that there are insect predators, and also more birds around to eat bug

              • > When I've done organic gardening ...

                When I build electronic circuits, I sometimes use wire wrapped. You'll never find wire-wrapped circuits in any store, so the fact that I use it is completely irrelevant to any discussion about which electronics people buy. When I build model planes at home, I use Dollar Tree foam board, gift cards, and popsicle sticks. You won't find any stores selling planes made of DT foam board, gift cards, and popsicle sticks, so that construction method is irrelevant to what peo

          • >Did you know that mint oil is a neurotoxin to wasps?

            That would be interesting, and relevant if wasps were a pest to important crops, and therefore mint oil could be used to protect the crops from wasps. As you may know, that's not the case.

            Well, I suppose it's indirectly relevant if you consider that wasps primarily feed in insects, some of which are crop pests. Therefore using the organic mint oil kills the wasps, which then can't kill the pest insects. In that way, organic mint oil increases the numb

  • by HuskyDog ( 143220 ) on Saturday May 26, 2018 @12:43PM (#56678892) Homepage
    The article seems to suggest that the plan is to use current herbicides in greater concentrations to overcome resistance. But, surely the entire point of the current herbicides is that they fiendishly designed to be poisonous to weeds but not to crops. Once you have a gadget which only sprays the weeds, you can use a less specific chemical which just kills any plant it is sprayed on and which presumably the weeds haven't yet had a need to develop resistance to. It also seems to me that a weedkiller which kills any plant can attack much more fundamental aspects of plant biochemistry and therefore be harder to develop resistance to.

    I already have weedkillers in my shed which basically say on the label "Kills all plants, only spray this on your weeds" and what they seem to be proposing is essentially the huge automated equivalent of me roaming my garden and carefully spraying any weeds I find.
    • The article seems to suggest that the plan is to use current herbicides in greater concentrations to overcome resistance. But, surely the entire point of the current herbicides is that they fiendishly designed to be poisonous to weeds but not to crops.

      It doesn't work that way. They are only MORE harmful to the weeds than the crops, and even then, only when you re-engineer the crops to be resistant, and even then, only until the weeds become resistant (which apparently only takes twenty years.)

  • by Solandri ( 704621 ) on Saturday May 26, 2018 @12:46PM (#56678900)

    Weeds are already becoming resistant to such glyphosate-based herbicides after "more than 20 years of near-ubiquitous use," reports Reuters.

    Note that this invalidates most of the early court rulings in favor of Monsanto, which were made under Monsanto's assurance that plants could not develop resistance. And thus any crop which could survive spraying with Round-Up must be from stolen Monsanto seed. This shifted the burden of proof in Monsanto's favor. The farmer had to prove they were innocent and the seed got on their land accidentally, instead of Monsanto proving they were guilty because the farmer planted it deliberately.

    • You think that plants in the fields of different farmers around the world all evolved the exact same gene sequence independently? And the farmers just accidentally discovered this and then started spraying them with roundup?

      That's adorable.

    • by HiThere ( 15173 )

      I don't think logic works for legal precedents.

  • by Tablizer ( 95088 ) on Saturday May 26, 2018 @01:38PM (#56679152) Journal

    This is the kind of research that should be subsidized because of the potential health benefits of reducing herbicides in our food.

    Market forces alone will not entirely "care" because they are weighing costs and profits, not long term consumer health.

    True, it may make things easier for organic farmers, but if weed bots get cheap enough for regular crops, then either they are cheaper than herbicides, or herbicides can be banned because weed-pulling robots would a sufficient alternative.

  • AI-Enhanced Headline Writing Robots Frighten Tech-Blog Editors

    A Swiss company called headRobotix is betting the tech-blog industry will be willing to welcome their solar-powered headline-writing autonomous robot, in an effort to reduce the use of lame editors by up to a factor of 20 and perhaps even eliminate the need for clickbait-enhanced headlines entirely.

    The 'see-and-write' robot goes from story to story, visually differentiating the actual information and FUD, and writing the headlines selectively an

  • by FudRucker ( 866063 ) on Saturday May 26, 2018 @02:39PM (#56679388)
    and instead of pulling the weed it just burns them to the dirt, it might be quicker and the dead weed can just sit there and decompose back in to the soil
    • Now all that’s left is to develop an on-land survival and mobility suit for the sharks who’ll be wearing those lasers...

      What could possibly go wrong?

    • Doesn't work. I tried to make a laser grass cutter once. I failed dismally, and still have the retinal damage*. I did learn that plants are surprisingly hard to cut with lasers, due to their high water content - all the laser does is burn the very surface and start boiling off the water, thus keeping the plant from getting hot enough for the burn to penetrate.

      *I was manning the camera and inadvertently caught the reflection.

    • and instead of pulling the weed it just burns them to the dirt, it might be quicker and the dead weed can just sit there and decompose back in to the soil

      Doesn't kill the roots. Many weeds are notoriously hard to kill.

  • by h8sg8s ( 559966 ) on Saturday May 26, 2018 @03:44PM (#56679604)

    This disruption couldn't happen to a mode deserving bunch than the likes of Monsanto. The real agri-revolution will happen when everything from in the agricultural production cycle (plow/seed/weed/pick/sort/transport/etc..) is automated. In that world, what will farmers be transformed into?

  • by pubwvj ( 1045960 ) on Saturday May 26, 2018 @04:04PM (#56679688)

    I already have weeder robots on our farm. They eliminate the need to use herbicides and pesticides because not only do they pull the weeds but they also kill rodents and insects that are problems for our crops. I have the versions called D.U.C.K.S. and C.H.I.C.K.E.N.S, - They work great.

    Seriously. Properly timed rotations of ducks and chickens through corn, brassicas, tomatoes, peppers, pumpkins, sunflowers and other crops do wonders to remove the weeds prior to planting and after the crops have grown to about 12" to 18" in height.

    Prior to all of this we use the biorobotic tilling machines called P.I.G.S..

    As an added benefit all of these robots excrete organic fertilizers that are specially formulated to make our crops thrive. It's called S.H.I.T.

    Finally, at the end of the biorobot's service life rather than repairing them we disassemble them and pack up the drive components to sell as high quality organic meat. What a win-win situation for farmers, consumers and the environment! Of course, pesticide producers are rather P.I.S.S.E.D. but that can be used for fertilizer too as it is high in beneficial nitrogen need by crops.

  • Robounfillers (Score:3, Insightful)

    by Impy the Impiuos Imp ( 442658 ) on Saturday May 26, 2018 @04:23PM (#56679774) Journal

    Soon cities will profit selling off the right to rip open landfills with robots and recycle it all.

    You read it here first!

    Actually, if you had been paying attention you read it here 15 years ago here first. I usually got downmodded saying it was silly to worry about landfills.

  • Sure they'll come back up, but if there's a pass of this machine every 3 days, it'll eventually kill it.

  • The pesticide industry will do just fine even with weed killing robots being deployed. There are plenty of other chemicals to make and plenty of other facets of agribusiness they can profit from. Even if sales of roundup permanently went to zero this afternoon the company that makes it would be just fine.
  • Now do the same thing for fire fighting. Have a solar robot stationed and patrolling a given area. When a fire ignition is detected, it sprays water to extinguish the flame at the source.

"Once they go up, who cares where they come down? That's not my department." -- Werner von Braun

Working...