Thanks to the wonders of neural networks and machine learning algorithms, it’s now possible to do things that were once thought to be inordinately difficult to achieve with computers. It’s a combination of the right techniques and piles of computing power that make such feats doable, and [Robert Bond’s] ant zapping project is a great example.
The project is based around an NVIDIA Jetson TK1, a system that brings the processing power of a modern GPU to an embedded platform. It’s fitted with a USB camera, that is used to scan its field of view for ants. Once detected, thanks to a little OpenCV magic, the coordinates of the insect are passed to the laser system. Twin stepper motors are used to spin mirrors that direct the light from a 5 mW red laser, which is shined on the target. If you’re thinking of working on something like this we highly recommend using galvos to direct the laser.
Such a system could readily vaporize ants if fitted with a more powerful laser, but [Robert] decided to avoid this for safety reasons. Plus, the smell wouldn’t be great, and nobody wants charred insect residue all over the kitchen floor anyway. We’ve seen AIs do similar work, too – like detecting naughty cats for security reasons.
What is this, a laser for ANTS!?
i could use one for skeeters, but the size of the skeeters in this part of the world you would need an airborne laser system to take them out. giant bloodsucking helicopters.
could do like we do down here and use a .22, only bounces off one out of every 5 or 6.
Wasn’t there a project do one for mosquitos?
IIRC it was a Bill Gates funded project to combat malaria by killing the females (?). Cameras and an illumination laser were used to detect a female mosquito by the distinctive frequency of the female’s wingbeat and determine its location. A kill laser then smoked the target insect. The lasers were setup as a kind of fence, zapping only those that crossed the barrier plane. Seemed like an awfully complicated mosquito net.
It wasn’t Bill Gates funded, but it was run by a former Microsoft CTO.
Apparently the company is still going, but they’ve shifted their focus away from the developing world and towards things like outdoor events in rich countries. I guess they finally realized what everyone else pointed out immediately: expensive, fragile, hard-to-maintain solutions like lasers are not well suited to solving problems in developing countries.
Shifted their focus
how else were they going to get the B&MG grant money?
Yes, this was done by Nathan Myhrvold at his Intellectual Ventures Lab. I got to see the prototype in action and it was impressive. The system tracked mosquitoes all the way down a long hallway and zapped them with a laser.
Didn’t this used to be just a magnifying glass?
“Twin stepper motors are used to spin mirrors that direct the light from a 5 mW red laser, which is shined on the target”
By making this sentence passive, you’re opening a whole can of worms regarding shined vs shone!
Seems to be an American thing, they seem not to use the word “shone” for un fathomable reasons.
probably because we would have a hard time distringuishing it from shown. Was the laser shone onto the ant or was it shown to the ant
I was promised zapped ants, and all I got was lousy flashing lights! ;)
The gardener in me really wants this for killing squash bugs and vine borers. Nasty little devils.
If they have natural enemies, use them;
if they are very fast, use chemistry or physical barriers;
if they are slow, use mobile robots with grabbing hand(s);
if they are neither slow nor very fast, use lasers.
Don’t tell Greenpeace about this cruelty…
Machine learning and neural networks needed to track moving objects?
Don’t tell the military , they have had tech to do this long before these these buzzwords came about.
There is always at least one person like this in the comment thread. IT WAS A FUCKING LEARNING EXERCISE. Your easy snark is like complaining about the guy who built the 6502 CPU out of discrete FETs because he could have done it far easier with one IC.
Yes. A learning experience. That’s the bulk of what you will do in this hobby. The “what if” projects are whupped together, tested, a week of fussing with it at most, then (if smart) torn back down to go back in the bin. I do not maintain this last discipline, about half are just tossed in a box and MIGHT get scavenged if I remember those parts are on it.
There is a big difference between hacking an experiment to just try, and building something to keep. Hackers do both.
Hack on!
The geek in me is intrigued by the technical challenge. My pragmatic side knows it’s not a practical way to rid oneself of an ant infestation. The inner Buddhist in me kind of recoils a bit… it’s sort of on a level with ripping the wings off of flies, burning sundry insects with a magnifying glass…
Not approved.
This is neither practical nor efficient but cruel (under the assumption invertebrates process pain similar to vertebrates).
Absolutely!
in that case, have you considered the level of chemical and biological warfare in which we’re engaged with these little monsters?
Coming to a TSA near you: ant profiling.
A similar construction is used to zap sea lice: https://spectrum.ieee.org/semiconductors/optoelectronics/licehunting-underwater-drone-protects-salmon-with-lasers
To the “but that’s cruel” comments:
What are you talking about. He just shined a light at them. He isn’t burning them! Besides if he was doing anything that hurt them we would probably see some sort of obvious reaction each time the laser hits the ant. I don’t see that in the video, it just looks like an ant being an ant, not caring at all.
To the “but I wanted the ant dead” comments:
Add one pet dog, cat, etc… who likes to eat bugs. It will figure out soon enough what those red flashes in the kitchen are and what to do about them! That sounds a lot better to me than laser burns all over the floor! Might even save some money on pet food.
I’ve wondered about something like this for detecting and killing weeds. Have a robot that drives around the lawn, identifying clover, dandelions, etc., and lasers them. I think identifying weeds among grass might be a pretty good challenge for any sort of computer vision algorithm.
The movie Andromeda Strain (1971) depicted an automated laser for rodent control in the secret laboratory’s service tunnels. In this scene, someone is being shot at by the system: https://www.imdb.com/videoplayer/vi2307784985