We knew it might happen sometime, and now it has..
"The Rise of the Superweed"
Remember 'Roundup' the Glyphosphate chemical developed by Monsanto. Roundup killed all the weeds on farmers fields, allowing him to produce more crops per acre, year after year.
Roundup eliminated all weeds, greatly reduced plowing, and therefore labour, and also wind erosion.

Just as the heavy use of antibiotics contributed to the rise of drug-resistant supergerms in humans, American farmers’ near-ubiquitous use of the weedkiller Roundup has now led to the rapid growth of these tenacious new Superweeds.
Superweeds are springing up---- not as many as clogged fields over 20 years ago....not yet.
But these Superweeds cannot be killed with standard weedkillers. And they are on the increase, can only be killed by spraying the field with a very toxic new modern weedkillers.
Higher food prices, lower crop yields, extra weeding and ploughing will add to food prices, as well as the pollution of land and water.
With the rise of the Superweed crops might soon have to be checked not only for toxicity, but also washed before processing for food.
Below is one farmers story. He is not the first farmer to suffer, nor will he be the last.
Nature is girding its loins, and I fear is about to hit back with a vengence.

http://www.nytimes.com/2010/05/04/business/energy-environment/04weed.html?src=me&ref=homepage

Mike
Isnt that the way monsanto works ?
they design products that eliminate their competition so
they have complete control.

then they design new more powerfull products that will kill
what they designed earlier.

this way no chemical company can out perform them.

they create a situation so that they can provide the remedy.

Isnt that the way monsanto works ?

Best ultimate solution to the problem is to develop mass-producible (thus cheap) mini-robots, perhaps solar-powered, that roam fields and pluck out any plant other than the crop, while they are yet very small. Thus all agriculture will be organic.

Warren, that's the best use of technology I've heard of in many a day. It would be even better if the mini-robots were assembled by a staff of robots. The whole process would be mechanized, and require no more labor than a manager to supervise the assembly 'bots. Thus putting more people out of work. Oops, there had to be a downside.

Thank you Amaranth Rose, that's a good idea about being made by robots, or at least mostly automated-- every field would need thousands of them, moving slowly like beetles to conserve energy, and small enough to easily move between the crops.

They could be used until they wear out, when they signal a problem and another one nearby carries them out to be recycled. At harvest time a universal signal would have them clear the field. Even though they might last a long time they would still need to be quite cheap.

Another advantage-- because they are machines rather than living things they would not be tasty to birds or other animals

Great Idea Warren

this would be the next step in industry removal as agriculture
is our only remaining large scale industry.

and if we can compete with the extremely low wage countries
using robots we just might be able to maintain our leadership
in agriculture.

I know that the solar industry is already using robots to clean
the solar reflector pannels in solar farms , so why no use them in agriculture also.

these robots would mean many high paid high tech jobs
and since shipping is so expensive that would mean that
many new robot factories would pop up to fill the need
for these robots.

heck when you look at a modern tractor these days you kind of wonder why they even need a driver in them , they could even be
opperated by a robot, they already have gps in them.

just program it and let it go.

international harvester already uses robotics to do most of the construction of the combines they produce in the U.S.A.
not in some foreign country , so at least they are wise
to whats happening , and they know that to be competitive in this
world of extremely low wages you must use robotics.

Wow! What a great idea. Sure it can't be one extreme or the other, but robots could be an extension of the farmer's hands and other senses. Robots could monitor soil moisture, nutrient status, and even provide some watering services while they are also culling weeds. Some of that information might be available soon via satellite-mediated remote sensing of the farmer's fields, but maybe the two methods could be combined; with robots receiving commands from the satellite (or providing calibrating information for the satellite).

Robots in Agriculture! Of course! :headslap:
There are all sorts of possibilities that evolve from that singular idea. I think these qualify as synergized solutions, the best kind. I "second" Paul's post. Good going Warren!

p.s. ...and Thanks Mike! for the very neat article on evolving resistance.

[quote=Mike Kremer]

Gosh, I had no idea you guys were so interested in Robots.
If it had'nt been for Amaranths tongue-in-cheek posting, I would have renamed this post the "Rise of the Robots",
..Or the demise of humans. (thru lack of jobs, food, money and sheer boredom).

Thank you guys for the positive comments on my suggestion. Got the idea from those cheap little vacumn/dusting robots that you turn loose in a room and thought, why not a random search through a field.

But if the crops were neatly spaced in neat rows they wouldn't have to even discriminate between different plant forms, but just go straight down between the rows and remove anything growing that was out of place, making the technological challenges simpler.

Yea awesome idea. But you stole it from me :P While I was working as a farm weeder I told my boss that I wanted to make a robot to replace our jobs, and she just laughed!! Weeding really is a simple and mindless task.

Another advantage of robots over weedkiller is sometimes the problem isn't weeds, it's other crops that have contaminated the field. So they're already effectively Roundup proof.

somebody beat us to it , there already here , farm robots

you need to learn how to time travel , it seems.

but I dont see any micro robots yet , and they should be much cheaper.
especially if they are built mostly by robots and just supervised , programmed
and serviced by human beings as amaranth sudgested.

Obviously Paul they just need to be improved and perfected, else they would already be in every field in the Midwest!

Warren, there has to be some discrimination. Weeds sometimes grow IN the rows, and they have to be deleted somehow. So the pattern recognition software would have to take that into account. But it's not too difficult to tell a weed from a crop plant. Okay, maybe it's hard to tell shatter cane from corn or milo (sorghum), but that would be a worst case scenario. For the most part it's pretty easy to tell what is a weed. Of course, you wouldn't want your corn programmed robots invading a soybean field and mistaking the soybeans for weeds, but a little creative engineering would make that a non-problem. It's all in your focus.

Sounds like companies could start working on that now, since I believe we have that technology— just have to put it all together.

Warren
if you want to pursue this idea , or to just develope a web
based application for the robots using gps , or a web based community , or any
other web based application try the below link . you can get anything you could
possibly need over a three year period to accomplish almost any goal.
and it only cost $100.00 us at the end of the 3 years.

there are also companies that are willing to help you
start up your buisness with $18,000 start up funding.

so have a look.

P.S.
Microsoft Robotics Studio is also there , the full
version not trial software.
all the software and opperating systems are full retail version , of course there are betas available.



Microsoft Biz Spark Program , all the software you could possibly need to start

Thank you Paul— I would probably need an investor and a robotics engineer, and on top of that I am a city boy who knows little about farming. But am toying with the idea of venturing forth with a website of some sort to start things off. Thanks for the ideas to get started. —Warren

Looks like there's already been a lot of work done--

"Autonomous robotic weed control systems hold promise toward the automation of one of agriculture's few remaining unmechanized and drudging tasks, hand weed control. Robotic technology may also provide a means of reducing agriculture's current dependency on herbicides, improving its sustainability and reducing its environmental impact. This review describes the current status of the four core technologies (guidance, detection and identification, precision in-row weed control, and mapping) required for the successful development of a general-purpose robotic system for weed control. Of the four, detection and identification of weeds under the wide range of conditions common to agricultural fields remains the greatest challenge. A few complete robotic weed control systems have demonstrated the potential of the technology in the field. Additional research and development is needed to fully realize this potential."

http://portal.acm.org/citation.cfm?id=1343305

A little late replying, just saw this thread today...

I don't think any of us in the biology field are surprised by this occurrence; its exactly what you would expect given how evolution works. It was never a matter of "if" resistant weeds would evolve, but rather a matter of "when". And Monsanto isn't alone in this - pesticide-resistant pests & weeds are arising to nearly all major pesticides and herbicides.

Its inevitable; since its impossible to ensure that the enterty of a field is covered with a lethal dose of pesticide/herbicide, you'll inevitably have regions where pest species are exposed to sub-lethal doses of the agent. Any slight resistance will be selected for in these areas, leading over time to a population of organisms capable of handling higher doses of the agent. Repeat this over time, and eventually you select for a resistant pest species.



I doubt monsanto (and the various other pesticide/herbicide produces) planned this. Keep in mind we live in a world where most people don't "believe" in evolution, and even among those who "believe" in it, a large portion remain ignorant as to how it works.

Biologists have been warning, for decades, of the inevitable rise of resistance to products such as these - and we've been ignored. But what you see today is exactly what we predicted - antibiotic (and now, even sanitizer-resistant!) bacteria, pesticide-resistant insects, herbicide-resistant weeds, etc. Monsanto et al are not making money off of this - as their products become less effective, they loose sales. No replacement products are on the horizon, and given the higher safety and environmental standards of today, development of replacement products is prohibitively expensive.

Monsanto is a perfect example - how many new herbicides have they developed (new as in new compounds, not new mixtures of old compounds). The answer - zero (I think).

Basically, you're looking for conspiracies where none exist - human stupidity, not deviousness, is all that it took for this to happen.

Bryan

I think it's great. If weedkillers don't work it'll force those weeding robot makers into gear. Afterall it's probably only the existance of cheap weedkillers that's keeping them from rapid development and use.

Then of course there'll be the robot-resistant weeds that fight back with their fiberous, planty skill-saws and hammers :P

I have a cheaper option - use <insert currently troublesome group of illegal immigrants here for your particular local>. Probably cheaper than robots, although far less PC...

Boy, I hope that comment doesn't get me banned. But the point is that at todays prices, minimum-wage labor is likely far cheaper than robots.

But robots are waaaaaaay cooler.

Bryan

I 100% agree to this , as Im sure monsanto does also.
and Im also inclined to "believe" that monsanto along with
the other pesticide companies fully understand that they are
responcible for the resistant species.

just like BP is responcible for the oil leaking into the gulf.

farmers who can no longer grow crops due to the resistant species
that are the result of using these pesticides should be able to
recieve monetary loss recovery from monsanto and the other pesticide companies
along with those resturants that also are damaged by the pesticide use.

otherwise there is no end to what will happen , monsanto and the other similar
companies should be held responcible for their actions just like BP isbeing held responcible.

so when you think about it , as monsanto surely knows the damage they are
causing the conspiracy that you speak of is not conspiracy its fact.

of course getting them to admit that is another story.
and getting a biologist to admit to that is another story , like getting a oil
man to admit that burning oil is a bad thing for the environment.

when you think about it on a one time only basis , perhaps.

but hireing a single laborer for $7.00 an hour x 8 hours = $56.00

suppose you need to have at least 100 laborers durring the 8 hours

thats $5600.00 a day ---- that money is gone.

and they only work 8 hours.

but there are 7 days in a week

you have spent $39,200 a week ---- that money is gone.

a gps robot should be mass produceable for less than $200.00 each.

and they can work 24/7

so 8 x 7 x 100 = 5400 hours a week

vs the robots

24 x 7 = 168 hours a week each

5400 / 168 = 32 robots to work 5400 hours a week.

32 x say $300.00 each = $9600.00

the week is over and you still have your robots to work the next week.

think of the savings to farmers and the savings to the environment.

and maybe you could rent the robots instead of purchasing them.

and the provider handles the upkeep of the robots.




I think its a great idea and I think its a way we can compete with the
ultra low wages in foreign countries.

I don't agree here, for two reasons. Firstly, knowledge of evolution isn't hard to come buy. That inappropriate use of these products will lead to resistance is a well established fact, and it is upto the end user to take that into account and thus use the products properly. Secondly, the pesticide products I am familier with (I grew up on farms, and most of my family remain farmers) come with instructions on proper use - including on how to use them in a way that reduces the likelihood of resistance developing. I would argue that given the farmers should have this info at their disposal, it is therefore their responsibility to make sure they use these products appropriately.

The comparison to BP is simply wrong; a closer analogy would be the rise of antibiotic resistance. It isn't the antibiotics manufacturers fault this has occurred, but rather is the fault of the MD's who over-prescribed them, and the patients who misused them.

What you want to do is hold a manufacturer responsible for the mis-use of their product by a consumer. That's arse-backwards.

My family are largely farmers. They are not having problems with pesticide resistance, because they use the products properly. Why should farmers too stupid to follow the instructions on the bottle get a free pass? For that matter, why should the manufacturer be held responsible for their inability to follow the instructions on the bottle.

Bryan

You've got the costs of labourers down, although your $7/hr is unlikely. Most are paid by what they produce/pick; not by hourly wage.

But you've ignored the other half; the cost of the robots. I have no idea what such a unit would cost, but given that the simplest of modern manufacturing robots run over $10K each, and a farming robot would have to be far more complex (needs high dexterity to avoid damaging the crop, visual recognition systems - something I have some experience with, and trust me it ain't cheap or easy, must be mobile, etc), I'd be guessing you're talking about machines in the hundreds of thousands of dollar range (equivalent to a combine, btw, if your talking the sub-$250K range).

There is also the issue of multi-use. A laborer can pull weeds, pick produce, remove pests, apply fertilizer, water the fields, preform repair work, plant seeds, setup bird/shade netting, etc. Meanwhile, most farm equipment (and thus, presumably, most farm robots) can only do one or two functions. Ergo, your one laborer may need 2-3 robots to replace them.

And then you have operational costs (fuel, repairs, etc), maintenance, mandatory warranty work, and the wages of highly trained (and thus expensive) operators, etc.

And the last problem being the issue of farming being a seasonal activity - during which time your robots sit idle, but still cost you money (be it storage fees, winterizing expenses, or whatever).

One day I'm sure the technology will develop to the point where these things will be possible in a price range which is competitive. But given the state of robot tech today, and the relative inexpensiveness of seasonal laborers and students, we're a long ways off.

Bryan

the reason robots and especially small weed picking robots
are so expensive is they are not mass produced.

how much does it cost to build and deploy a communications satelight?

yet I can get tv , internet , telephone for apx $80.00 a month.

no where near the billions of dollars it cost to develope the satelight , build the rocket to put it in orbit or pay the salaries to all those involved.

but its the mass service contracts that pay the billions of cost.

just like the farming robots , but in mass production the small weed picking robots will be made of plastic , have photo comparison software in something like a cell phone sized computer
that takes a picture of the ground then determines the weeds position , then moves to that position and plucks up the weed.

it would resemble something like one of the remote controled
trucks or cars that you see selling in radio shack for around
$30.00 , it just doesnt have the photo comparison software or a
onboard camera array.

or a artificial weed plucking hand.

but with mass production those items cost would be greatly lowered also , just like the remote controlled truck and cars.

I said $300.00 because it would only be farmers mostly that would be purchasing the robots , not the general public , if this type of item would sell to the general public its cost might approach $50.00

In part. But robots are also fantastically complex, which adds to the price. Keep in mind, complex objects like cars and planes are mass produced, and yet they are quite pricey.



Yeah, but your not buying the satellite, nor are they mass produced. What you are paying for is permission to intercept the signals from one. Hardly a relevant comparison.



I think you underestimate the complexity of IDing a 3D object, separating it from the background, and identifying what it is. As part of my day-to-day job I write software which automatically analyzes 2D microscope images (which have a only a single colour, btw), identifies and classifies objects within those images, and preforms various forms of analysis and processing.

Despite the "simplicity" of the image processing I engage in, it requires a large distributed network of computers to preform that analysis. We're talking a total of 32 computers cores (plus an additional 8 that preprocess and oversee the whole process). These processes are far simpler than what would be needed to ID a weed verses a crop (which would require colour analysis, morphology analysis, the ability to ID a 3D shape given multiple possible viewpoints, be able to ID different developmental states, etc), and even given the massive computing power we have still takes hours to run. And you think a cell phone-level computer is going to be able to preform a far more complex tasks in seconds?

Also, keep in mind that the robot you propose isn't going to offer farmers much - its simply taking care of one task that a laborer could otherwise do while engaged into other tasks. If you're going to have a guy/gal walking down the rows of your crops planting, picking, running watering lines, tasseling, etc, you may as well have him/her pull weeds at the same time. Why spend hundreds (I'd argue thousands to tens of thousands) for a robot that does one task, when you're going to have to hire humans anyways, and they could do the same task as part of their day-to-day duties?

Bryan

geez, thats alot of work.

well if the robots are the ones planting the seeds , then
they know where the plants will grow (data storage).

everything green around that area could be plucked up.

so all you need is a color identifier.

a simple color number program can do that in a few lines of code.

ie...
pixel = 0
color = (20000 to 20100)
sub pick_color()
for pixel = 0-50000000
get image.pixel.color.value
if image.pixel.color.value = color then goto pluck
pixel = pixel + 1
next pixel
end sub

sub pluck()

load the image , scan the image with code allow a few hundred shades of green , compare
the value of each pixel , note the depth in the image by the base of the color (the weed color), plot its location , snap a seperate image from a different angle repeat the above , now you have the weeds location.

it just gets simpler , believe me.

just make sure you dont use green dirt, LOL

I could even see how you could have bean picking robots.
corn picking robots , it would bossom into a gigantic robotics industry.

set it and forget it.

its a good thing there wernt so many nay sayers around when everything was getting invented , or we would still
be rubbing sticks together to make fire to cook what we just killed , because there is no refrigerators or ovens
or lights , or anything for that matter.

try optimism.

Except that not all seeds germinate, seeds move within the soil, animals will move the seeds, not all plants grow in straight lines, and a weed can occupy the space "assigned" to a planet by out competing it - that's what makes them weeds after all.

May work for hydroponic operations; doesn't stand a snowballs chance in hell in a field.



Except for a few not-so-minor issues:
1) Many weed "species" are the wild variants or closely related wild species to the crops under production. This is why they are weeds - they closely resemble the crops under cultivation, and thus grow well under the conditions farmers create. Ergo, weeds often look pretty much the same as the crop. This is a major issue in canola and wheat farming, and is also a driving factor behind producing herbicide resistant crops.

2)Even in the same species of plants, the colour can vary greatly depending on a range of factors - differing intensity of sunlight can alter the production of protective pigments, some pigments are hydration and/or pH sensitive, disease often bleach colour, colour often changes at different stages of plant development, and all that's on top of good 'ol phenotypic variation.

3) The range of colours available to plants are limited - all land plants have the same fundamental 5 pigments, and their "specific" colour is merely due to the balance of those pigments. As a result, the number of unique colours is quite limited.

Together, those little details make colour-based identification far more difficult than you think. The chance of a crop having a colour - even a spectra - unique enough to separate it from weeds is pretty much zero. Both the limited colour options, combined with the colour variation in single species, precludes this type of detection in most cases.

To actually ID weeds you need to analyze morphology: the shape, size and spacing of leafs, the vascular patterns on those leafs, the branching patterns of the stem, etc (this is how plants are ID'd in the real world). And your algorithm needs to be robust enough to deal with damaged plants, stunted plants, wilting plants, and weed species closely related to your crop.

Its a tall order, far more complex then you seem to realize.



Hmm, I'll raise your supposition with a childhood spent farming and an adulthood spent developing computer routines to analyze morphological features.

That's a poker analogy, BTW. What I'm trying to say is this task is far more complex than you seem to realize - plants are not nicely colour coded for your convenience, nor do they play nice by always being clearly different from each other. I've got a lot of experience in using computers to ID objects based on morphological features, and if were as easy as you seem to think, I'd have been unemployed years ago...



Get on it then - if its so simple it should be no problem for you to develop said systems.

That said, the technology to pick beans, corn, etc, is already largely automated. All you really need to do for that is put a tractor with the appropriate harvesting equipment under GPS control. And do it at a price cheaper than having your kid drive the tractor...

Now an automated tomato, or pepper picker; that would be a thing of beauty...



I prefer reality, thanx. As someone whose made a career as a scientist (and I've invented a couple of microfluidic devices - maybe that qualifies me as an inventor as well?) I've learned that the quickest way to success is to ID the weak points in your idea, and attack them head-on. The granting agencies have forced me to also appreciate the cost-benefit side of planning...including knowing when an idea isn't financially tractable.

Your route leads to a lot of failures and disappointment. My route gets the job done - if, indeed, the job needs to be done.

Bryan

maybe you just read my last reply wrong , I never said that the color would be used to identify the plants , I just said if the color is green and not in the designated location it is plucked.

maybe I wasnt to clear about that part.

I myself have planted seeds and have noticed how they for the most part shoot straight up after the sprouts leave the seed.
possibly veering no more than 1/8 inch further from where they are planted...depending on the seed size.
unless there are rocks or some other obstacle in there way.

which pretty much covers your reply.

listen I understand how something such as this would threaten
your job , so I dont blame you for securing your future.

after all , why would farmers need biologist if they grew there own seed , or were capable of growing there own seed as is the norm these days because of biologist fixing seeds so that they cant.

People who are always suspicious of others are often guilty of the same thing themselves.

But I do agree with you that identifying weeds isn't that big a deal if it's mass produced. Sure the software will be difficult to design. But once it's done it's done. Hell my cheapy digital camera can recognize poeple's faces. This stuff isn't trivial but copying software is free, so the more you sell, the cheaper the development cost is per unit.

If it makes a few mistakes it doesn't matter. The existing use of humans for weeding is pretty unreliable too. Also no need for an expensive robot grabber to dig them out, just spray them, burn them, mow them or whatever.

No, you were not.



Then you obviously haven't worked on a farm. We plant seeds using a machine then places individual seeds in a perfectly straight line, with the seeds centered an exact distance apart (a surprisingly simple device; basically a set of tubes set around a drum. Seeds run down the tubes, into the soil, while the drum turns). And yet, by the time the seeds sprouted you had rows of grain 2' wide or more, with irregularly spaced plants.



<sigh>, going for the insult angle I see.

As my profile clearly states, I work in the medical sciences. Whether or not someone develops an automated farming device of any sort has no direct impact on my life what-so-ever.

But hey, ignore my experience in both the areas your referring to - your gardening skills obviously make you the expert...




Interesting world you live in, where the very people who've made it possible for farming to actually be able to feed the 6+ billion people on this planet are the bad guys.

Bryan

But the problem was never the software; given a few weeks I'd probably be able to come up with something that could do a good enough job.

The problem, as outlined earlier, is the issue of computing power - the kind of analysis you would need to analyze plant morphology on the fly is going to take a computer a little bit more powerful than a cell phone. If you want the robot to be able to tell weeds from crops at a reasonable rate of speed (i.e. so you could pick them faster than they grow back) you're going to need a fair amount of processing power; several thousand dollars worth at least - per robot. Plus GPS (or something else to keep them in your fields, and to ensure they go everywhere they need to), plus the cost of the robot itself, plus whatever fuel it uses, plus maintenance, etc.



Yes, but facial recognition is about as easy as morphological analysis gets - you're simply seeking a fairly uniform object (despite our preconceptions, the human face is fairly uniform, with eyes, nose and mouth set in a concrete spatial relationship). And keep in mind that your camera only determines if a face is there. It's not identifying what race that face is, if its a boy/girl face, etc - and those kinds of detailed determinations are exactly what would be needed to ID weeds verses crops.

And plants are much more diverse, morphologically speaking, than are people - even when you're talking about plants of the same species - making an even more difficult issue.

And don't forget, you do a lot of the work for your camera before it recognizes a face - like pointing it at a human, and having that human orientated in a fashion where the face is front-on to the camera. A weed seeking robot doesn't have that advantage; it needs to ID the weeds regardless of the preconditions; even if they are sandwiched between two other non-weed plants.



We don't use humans for weeding much at the industrial scale, so I don't think that comparison can be made. If plant production is your livelihood, you're not going to be willing to have much leniency. A 1% error rate is a 1% crop loss (ontop of other losses). Given the small margins on farming I don't think many farmers would willingly invest in equipment they know will increase the loss over that of the alternatives. As much as we hate them, pesticides don't cause crop loss, which is why they are so very popular.



Fire = bad; I've seen fields accidentally lit by a careless cigarette; a blow torch is out of the question unless you're farming something really wet. Spraying kind of misses the point - the whole purpose of the robot is to avoid chemicals (a laudable goal, IMO). Mowing doesn't get rid of the weed, only slows its growth...

Like I said, its a great idea, but there are a lot of hurdles in the way. And unless you can implement it in a way which is equally efficacious to current practices, and in a format which is financially competitive with the current practices, no farmer on earth is going to adopt the technology.

After all, who is going to buy dozens of robots, at hundreds to thousands of dollars a pop, when they can spend less on pesticide, spray it using their existing equipment, and experience less crop loss as a result?

Its a great idea, and I'd love for someone to prove me wrong on this, but its a far more complex issue that you guys seem to appreciate --AND-- you seem to be completely ignoring the economic angle.

Its like inventing a better mouse trap - sure you can invent better ones, but unless they work as good as, and are as cheap as, the wood/wire ones (3 for a buck at the local dollar store), no one is going to buy them.

Bryan

Hi ImagingGeek.

I also work designing processor intensive software, so I'm aware of the limitations of computers. And I've done weeding on farms - distinguishing Italian ryegrass from perennial ryegrass for example. That's why this whole robot idea excites me. Sure it won't happen straight away, but oneday I hope nobody will have to do such sub-human work.



Oh that's all you're talking about? That's nothing compared to the cost of building the mechanical parts without mass production. A GPS receiver is effectively free. I have a feeling you're trying to make the software problem sound expensive just because it's complicated. 15 years ago automatic number plate recognition was unheard of, now it's done routinely from moving police cars. Car-park guards can type your number plate into their computer and their cameras will find where you put your car. Addresses on letters are scanned faster than you can blink. Sure these may be easier problems, but they have some tougher cost/performance/portability requirements.

In weeding I think there are 2 separate applications which you're confusing. One is to replace selective weedkillers that you spray over everything. The other is to replace manual weeding. I don't know what this is like elsewhere but when I was doing it, I was usually picking out a single unwanted variety from a field of one or two OK ones. Choosing between 2 or 3 possibilities is surely much cheaper/faster than comparing to a database of thousands.

The whole farming thing gave me plenty of motivation to stay in school. There is a lot of boring, hard and repetitive jobs on a farm, and I'm happy to be away from that life (minus the farm gals, who are much more "earthly" than their city kin...).

I too would like to see a lot of this on farms, if anything, to increase the margins of my family who still work farms. It is not a profitable business, and its frustrating to see them work as hard as they do, for as little as they get.



But that cost does come down with mass production; the kinds of computing power needed though is currently expensive and will likely remain that way for years to come.

As I said, we do the same kinds of morphological processing here at work that would be needed to ID weeds verses crops. We make a point of regularity upgrading our hardware (its our single greatest material expense, after service contracts), but even with the most modern 8-core processors we're still needing to network several together to get the kind of processing speeds we need to do this stuff in a practical time-frame (and one not even close to real-time).

One day that kind of power will probably fit into my watch, but not anytime soon.



Quite the opposite - the software is easy, as morphological analysis is pretty old-school stuff these days. Like I said in my last post, give me a bit of time and I could probably write it. The issue is the kinds of computing power it'll take to implement that type of analysis, at a speed sufficiently fast to be practical, and at a cost which is competitive with existing options. If the robot can only process a few plants an hour, it'll be useless.



I'm thinking you may be a bit older than I Between GM crops and seed banks, farmers (at least up here) don't do a lot of selection on their crops. Even as a kid, which was a fair time ago (pre-GM, anyways), most farmers were switching to buying seed every year in order to avoid having to pay for the later kind of weeding you mention.



Not really. Its the actual morphological analysis which is computing-intensive. Once you calculate the morphological parameters, IDing the object takes milliseconds using fairly standardized cluster analysis routines (K-mean, or DBSCAN are the norm).

This is maybe beyond the range of this discussion, but in my own work I am trying to develop methods to automate the analysis of drugs and genetic mutations on certain cell functions that can be tracked with fluorescent tracers. The process is basically a 3-step one, separate the cell from the background, analyze the morphology of the cell and fluorescent tracers, and finally compare that morphological information to a huge database (currently ~50,000 cells strong, containing information on the effects of ~500 drugs/mutations).

The first and last step of this process takes less than a second each. The middle part takes anywhere from 10 to 15 minutes per cell. You can imagine, if a robot needed 10-15 seconds per plant to tell if it were a weed or not, the weeds would probably grow back faster than the robot would be able to remove them.

Bryan

Presumably the bots would be programmed to recognise key characteristics of the particular crop, then instructed to eliminate any non-conformers. It looks, from what you've all said, that were talking exclusively about visual ID. Would there not also be a detectable chemical signature for a given crop type?

And maybe our weeding bots could compare spectral reflectance:

"For weed detection in cultivated crops, two interrelated general approaches have typically been used ...The first is to detect certain morphological differences between the crop and weeds...The second general approach is based on differences in spectral reflectance...Despite the different problems encountered thus far in detecting weeds, some researchers argue that the spectral characteristics of plants are sufficient to differentiate plant species without introducing geometric complexities."

http://crop.scijournals.org/cgi/content/full/45/2/477

Nope just a couple of years ago. Actually I don't know what our clients were up to. Maybe they were making the seeds in the first place??



OK. So hey here's a way round it. Drive over the field recording everything. Then go home and analyse it at your leisure over days/weeks. Hopefully the plants are still where you left them when you go back to kill them.

At 10mins per plant you can do 1000 per week. Hmm that's certainly slow but maybe the processing could be done offsite at some dedicated facility and have it finished in a day.

How much would that processing cost?
Suppose 1 plant takes 10 minutes at 1 GFLOPS
That's 6e11 floating point operations.
Some efficient computer can do 540 MFLOP / J = 2e-9J/FLOP
So electrical energy use for 1 plant = 6e11 * 2e-9 = 1000J = 0.0003 kWh
@ $0.1/kWh
= 0.003c per plant
* 1 million(???) plants in a field
= $30/field
= Cheap!!!!

Multiply by 10 for actual cost of a computer service -> $300 per million plants, and add a bit for the "dumb" robot which might even be an attachment on the farmer's existing GPS guided tractor.

But that still doesn't change the time it takes to anlyze morphology, which is the slow step. Whether your database for comparison has 1 entry to compare to, or thousands, won't affect processing times much. DBSCAN and K-Means are very fast and effective clustering algoriths...



Possibly, although our technological ability to do that kinds of work is pretty limited - hence why dogs are still the preferred way of doing those kinds of detections. And I doubt it would work in the cases like canola farming, where the major weeds are simply wild strains of canola.

Bryan

If it works - great (spectral analysis is pretty cheap, and fairly light in terms of the computing power needed). But as your article states, even among the real experts that method is debatable. I imagine (having no knowledge of their field) that their issues are the same as the ones I brought up - variation in the same species, limited variability among species, and the fact that many weed species are the wild variants of the crop.

Bryan

I don't think that'll fly - weeds grow fast (other wise they wouldn't be weeds), and you usually want to get them early in the growing season so they don't crowd out your crops.

Besides, if you're going to spend time doing what you propose, it would probably be faster to just walk the field and pull weeds. And since human labour is expensive, faster = cheaper; plus no robots are computers are needed...

Bryan

if you know its comming , then you could start it.

if its something you like , then its not really work.

you seem to know how its done and the web based application
you mentioned would be a great idea as the cloud could perform the computations.

every farmer could allow a percentage of there processing capabilities to be used by the cloud , and you can build
the cloud yourself in an application that is downloaded and installed on the farmers computers.

there is probably plenty of funding available for something as this.

try the go big network , and remember that Microsoft has the biz spark program where you have access to every program that Microsoft sells.

I don't think farmers want to ctrl-alt-del their robots every 1-2 hours

Bryan

Quite possibly. There's only so much useful speculation to be done here regarding detection. But apart from the detection issue, there are other problems. These little guys would need to be capable of rapidly traversing soft and uneven terrain, and of killing weeds by some as yet undetermined method. They would also need a sufficient power source. Some people, sometime - my guess is quite a few years from now - may do the R & D and iron out the wrinkles using a series of prototypes. For the time being, this critter looks very bulky and prohibitively expensive; but I can see it coming. It's an idea too good to ignore.

Haha yea except a couple of important hurdles:
- I don't know it's coming, I just hope and try to reason that it must be.
- I don't have any business sense.
- I don't have a clue about image recognition.
- What I already do is so fun it's hardly work either

Yea this cloud thing might be ideal for it!! As long as the algorithms can be paralleled then you could bang out a field as fast as you like, or even in real time no matter how CPU intensive they are.

Exactly - if it were easy and affordable, it would have been done already. We've seen some gigantic leaps forward in a lot of technical areas of the past decades, and with time robots like this will be made with cheap, off-the-shelf-hardware.

Today, unfortunately, they are not.

Bryan