h a l f b a k e r yWhy on earth would you want that many gazelles anyway?
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
Holding weeds in check is a rather big problem in farming, usually solved with aggressive chemicals, or the combination of even more aggressive chemicals with crop-plants genetically altered to survive those.
The genetic modifications needed for crops to become resistant to plague, chemicals, etc.
are rather extensive, and often engender fear about those modifications escaping to weeds, or having unknown influence on other organisms in general.
My proposal needs genetic modification of the crop-plants, too, but on a far more bening level, that is already well-studied, and generally under control : green fluorescent protein, GFP.
Robots today are very fast, but to have them help in farming they would need to be able to make a difference between crop and weed, which they currently can't - the image recognition techniques are just not that good, and in many stages of plant developement it is simply impossible, even for humans.
But if the crop plants would have a GFP (not necessarily green, different colors could code for different crops), a robot would be able to weed out any other plant fast and efficient. The weeds could either help power the robot, or could be mulched.
No chemicals would be needed against weeds (as weeds are plants too, the chemicals going after the weeds always were tricky, as they had to exclude the crop from the slaughter), and as even the fastest-growing weed takes some time, the robot would not need to be supersonic.
Close...
https://www.youtube...watch?v=0lNFJBfZKAs Instead of herbicide a scratcher mechanism could be utilized. [whatrock, Jun 14 2019]
[link]
|
|
It just occurred to me, given the precision of modern
planters, it would probably be possible to produce a robot
that pulls or cuts anything that isn't growing within the
probable growth radius of each seed. This won't help if you
end up with a weed within a half inch to an inch of where the
seed was planted, but should deal with the rest. |
|
|
I thought of this idea yesterday while NPR was doing a story on
"superweeds"; weeds becoming herbicide resistant due to
overuse. |
|
|
But you wouldn't need dye markers. |
|
|
Shape recognition is getting pretty good. |
|
|
There are only six or so common weeds. Easy to program in the
shapes of the crop, then the weeds. |
|
|
Use GPS to guide the FarmBot (copyrighted) along the rows and
slice off the weeds subsurface. |
|
|
Somebody made a tomato-picking robot over a year ago that drives
itself, recognizes the fruit, determines if it's ripe, and gently picks
it. |
|
|
How much easier it would be to kill weeds than pick a tomato ! |
|
|
The FarmBot could have a solar panel and run all night on
batteries. |
|
|
Given how much herbicides cost the 'bot would pay for itself in
two years. |
|
|
I don't know how much I like eating fluorescent food, but this is an interesting take on a real problem, so enjoy a tasty bun.(+) |
|
|
Most farms I've seen, the crops are planted in a very orderly fashion... why not just "if it isn't where it's supposed to be, it's a weed", with occasional human (or tele-operated) inspection. |
|
|
Why not just fake the automation and make the system a giant mechanical Turk? Sell it as "fully automated" but then release a "Farming Simulator" with "astounding graphics and physics!" |
|
|
Then have people sitting in their homes drive the robots around pulling up the weeds for points that they use to buy digital hats. Farms save money, programmers make money, consumers get imaginary points. Wins all around! |
|
|
I think weeding robots are now either baked or at least in
trials. They use image analysis (shape, size and spectrum) to
identify weeds and then either zap them with a laser or apply
topical glyphosate. |
|
| |