When we first heard of [Ildar Rakhmatulin’s] plan to use OpenCV on a Raspberry Pi to detect mosquitos and then zap them with a 1 watt laser, we thought it was sort of humorous. However, the paper points out that 700,000 people die each year from mosquito bites — we didn’t verify that, but according to the article that’s twice the number of people murdered each year. So the little pests are pretty effective assassins.
It looks as though the machine has been built, at least in a test configuration. A galvanometer aims the death ray using mirrors, and with the low power and lossy mirrors the mosquitos can only be a small distance from the machine — about a foot.
Even so, the paper claims they could neutralize two mosquitos per second. We wonder how many of them survived but were blinded. There were several different detection algorithms in Python but even the best algorithms didn’t track 100% and the actual kill rate of mosquitos was quite low, topping out at 15%.
Clearly, this has some work to do, but if you decide to tackle it, the research will be invaluable. There was talk of using a different camera lens to get a larger volume of detection and, of course, a more powerful laser. If the tracking algorithm could be pushed to a smaller controller, the system could be light enough and power efficient enough to fly on a drone. However, we were unclear how you’d protect non-mosquitos from being hit with the laser of death. While a 1-watt laser might not kill you, even a 1 mW laser can produce effects on your eye greater than staring into the bright sun.
This could be more human than the last method we saw for exterminating the pests. Of course, as any time traveler will tell you, the best time to stop an assassin is before they are born.
The problem with this isn’t the laser. The issue is seeing the mosquitos against any background that isn’t white, and at a distance of more than a few feet. I built a turret that could easily point at a mosquito https://hackaday.io/project/165350-anti-drone-gun-turret, the problem is finding them in a real world environment. I think about it a lot but I still haven’t found a good solution. I think some narrow angle lidar might be the best bet.
Only if there was a way to see things without shooting photons at it…
The “ultrasound radar” type detectors usually tend to have range only up to 10ft and poor angular resolution.
unless they are in an array
True, hit ’em with those phaser beams… tending to ramp quickly in size, expense, and backend number crunching according to how accurate or fancy you want it.
There is a sophisticated, ultrasonic, hunt and destroy system for mosquitos, but it’s just bats.
Not that this such a large drawback. If there would be a product that would defend small terraces or balcony’s in the developed world this might work wonders for future development of the technology.
Passive detection: Listen. Triangulate. Aim. Fire.
Use the pesky buzzing of a mosquito. With a set of three microphones placed in a triangle, listen to the particular frequency of the mosquito’s wing beat. Delay in arrival of the signal on each of the microphones gives away the location, not the distance. So the laser should be in the middle of the three microphones. Maybe separate the mics with foam walls between them to also use signal amplitude for aiming. But then, that’s just the theory of sound ranging.
That’s right, that’s right!
https://medicalxpress.com/news/2020-05-lidar-track-mosquito-africa-combat.amp
Tried multiple colors of illumination? They might practically glow in one or another, they always look very white on security cams with active IR illumination.
UV light filter+emitter maybe would help, since a lot of insects have glowing spots when exposed to UV light. Then OpenCV wound do the rest?
Enclose the system in an 18″ cube with baffled openings and white internal walls.
Place something in the cube that would attract mosquitos.
The system would thus be eye-safe, would ensure that the mosquitos are within killing range, and the space could be tuned for optimal CV performance.
excellent idea, we could use UV light, and then throw in a two metal grills, one of high voltage potential and another at relative ground, and then we can ditch the laser/camera/things :)
You could sell a lot of those, I bet.
I totally agree to your idea for killing mosquitoes.UV light is intended to attract mosquitoes.
If you’re attracting them to a closed box then just use one of those electric zappers.
You also want to be careful attracting them you don’t attract them faster than you kill them worsening the problem!
You might have a box full of dead mosquitos but now even more bites!
How about an array of microphones that would detect that annoying buzz?
this!
Yes, I thought it was done in Florida, and could tell the different insects by their audio signature, in an attempt to protect the pollinators….
i did it , you can read it here, detect approximate position by audio and after that exact position by camera
https://www.researchgate.net/publication/349226713_Machine_vision_for_low-cost_remote_control_of_mosquitoes_by_power_laser
The original version of this that was discussed a few years ago used a secondary non-harmful targeting laser and a high speed cameras that could identity the specific wingbeat frequency and not kill anything but mosquitos.
i affraid it was to expensive, i saw video
Thank you so much, please this paper
https://www.researchgate.net/publication/349226713_Machine_vision_for_low-cost_remote_control_of_mosquitoes_by_power_laser
I tried to solve this problem with audio sensors, telephoto lens and deep neural networks, what you think about it?
so develop the target with a platform to have an included white background, align the bait or your favorite UV light (within range) and fire away ;)
Wouldn’t it be easier to put down say a mirror or sometihng with honey on it or anything that attracks flies? and point the turret about a foot away from it??
You can put before the camera a strong IR light because mosquitos reflect IR light so they will glow u can see this in security cameras this way u only tell the image processor just to seek for glowing small objects, that worked for me
Illumination with IR led array + IR filter on the scope wouldn’t work? As sneaky as the muffs are, they do reflect light…the CV could surely spot glowing blobs even against no backdrop. Just a thought, I plan to dive into this project as soon as I can squeeze a couple more hours into my day. Cheers
Do mosquitos emit a frequency?
If so, hunt the frequency.
IIRC, it’s only the male (non-biting) mosquitoes that emit sound.
Unless a female is trying to land in your ear, you won’t hear her.
There is already a device that zaps mosquitos. It’s called a candle.
We need to scale it up… a lot. oxyacetylene torch has both UV and high temp, but is a little pricey to operate for a long time.
When I was living in the bush in Queensland we had open flame carbide lamps that would be surrounded by dead insects in no time.
I believe they are attracted to the carbon dioxide produced by the flame as they believe it to be the exhalation of an animal breathing. I had a device that just had a purple/blue led and a fan inside that attracted the mosquitoes and sucked them into a mesh enclosure. The time in there dried them out. I never had another bite after I started using that.
or flamethrower
Malaria isn’t the only disease that mosquitoes transmit, but it’s the most lethal. Many vaccines have been created in the past, but the parasites always adapt.
One good that’s coming out of the pandemic is more brilliant people figuring out new ways to make vaccines.
Someone has used mRNA to generate a vaccine for malaria that doesn’t try to present the whole signature of the protein, including the part that constantly mutates. They designed a much simpler version that only mimics the protein fragment that doesn’t mutate.
Vaccines are a whole lot cheaper and easier to deploy than skeeter-lasers. Just not nearly as satisfying.
Most people think of Africa as being the only place to worry about mosquitos, but the little bastards are spreading West Nile Virus, Zika, and St Louis Encephalitis in North America.
Also a worry is if ticks can spread Lyme disease, couldn’t mosquitos biting same hosts do that also? Then one also could be concerned that prion diseases of deer could get transferred and turn into human CJD or similar.
interesting idea, thank you
It would be a lot easier to let people spray the inside walls and ceiling of any structures with DDT every couple of years.
maybe, but the problem still exists
Yeah, we could do that. We could nuke every square meter of the surface of the planet too. Who needs a biosphere anyway?
I don’t believe you can vaccinate against parasites. Malaria is caused by parasite.
It’s a single celled foreign body, it should be possible to provoke an immune reaction.
I’ll just leave this here: https://en.wikipedia.org/wiki/Mosquito_laser
I see that it detects female wing beat frequency of around 456Hz… now that would be good for leveraging passive detection… though I do also wonder what would happen if you blasted them with high intensity 456Hz, whether you could cause resonance to damage their wing muscles or something.
please make a mosquito disco stage box. they enter, they dance to death.
But are we neglecting the simpler, equally effective, solution?
https://i.redd.it/4po1dxcdw8yx.jpg
LOL!
(now I’m coughing from laughing so hard!)
Why didn’t I think of this before?? It’s so obvious now that I see it!
I still remember that one from some 10 years ago.
It was also mentioned on Hackaday.
Apparently it never got further then a few prototypes.
And if the hardware costs around EUR50 and if it really works it would be hard to believe it would have stayed that way.
I recall something similar. Wasn’t it sponsored by Bill Gates charity?
exactly!!! Bill & Melinda Gates Foundation, but I did not find a technical explanation
Often, Bill and Melinda don’t have technical explanations, as they seem to follow scams and donate monies assuming everyone is legitimate. Back in the 1980s, a process that worked in a lab, and failed to scale up, cost the county of Los Angeles millions of USD. Last year, I saw the same process repackaged and offered to the B & M foundation, to supply free energy to Africa. Of course, they gave up the money, but the process still will not scale. The process: Carver Greenfield, killed a few workers and wasted millions.. Go figure.
Probably the biggest problem would be patent troll Nathan Myhrvold.
Oh this is all a patent troll problem??? I always wondered why we don’t have the laser already, unless it’s because the explosion spreads the contagions everywhere.
Mosquitos use CO2 (+ body heat) to track hosts; could use a small CO2 cylinder to attract the critters into a dedicated (and safe for humans) kill zone
Bit of general purpose yeast and some brown sugar in water is a good co2 source in a pinch. Stick it in one of those ‘chopped the top of a soda bottle and flipped it upside down’ bug traps and watch the chaos. Satisfyingly effective. Also like the desiccated mozzies on a fan idea. Apparently works well and doubles as a gecko buffet apparently.
This would also make maneuvering the laser safer as it could be directed in a confined area.
Or use something similar to those salt guns. Although a lot of vacuuming will have to be done afterwards but the area covered will be greater.
“and doubles as a gecko buffet apparently.”
Goldfish love them too!
They have a distinctive pitch of their wings. Perhaps this can be exploited to kill them. Let them be picked up with a focused array of mics and pass thru the kill spot. No aiming, direct beam. I like to suck bugs up into a fan and filter with a light lure. CO2 is a lure but how to vent some and suck ’em up at the same time is a counter problem.
I want an anti-skeeter LASER hat, and for it to also blast chiggers, gnats, “noseeums”, and every other insect that wants to buzz around my head and dive bomb my ears. Some places a pesty insect LASER system would look like the “death blossom” attack in “The Last Starfighter”.
Distant memory of meeting someone who did this with a 30W fiber laser. Suppose at some point, there’s a trade of between higher power being better for zapping but worse for setting the room on fire.
A method used by my Finnish side of the family is to have a sacrificial person/bait that the mosquitoes prefer to feed upon. English blood seems particularly appealing to the little blighters. Unfortunately that’s the flavor blood I have.
30W of fiber laser would probably be lethal even to small birds…
The article talks about “neutralize” while Hackaday wonders if the critters are actually killed.
Killing them is hardly the point. Just slightly damaging a wing so it flies in circles would be enough, or indeed blinding, but they may be more dependent on IR sensor or detecting CO2 or something else.
This feels a little like déjà vu, I could swear HaD covered a mosquito tracking laser a few year’s ago.
Apparently my memory still works. :D
https://hackaday.com/2010/03/23/die-bloodsuckers-pew-pew/#more-22671
Here is the machine in non-lethal action https://www.youtube.com/watch?v=fwyMuwNYKvI
Surprised, didn’t expect that paper will be interesting for many people.
I’m sorry I didn’t make a video for the demo
But who is interested there is a video of the calibration process – https://youtu.be/2BKtM5cxOik – control yellow LED by laser
Now I replaced equipment to start to operate with deep learning
In fact, it is a really brilliant idea, which had already been formulated some ten years ago, as some comments have pointed out. It is a very clean way of reducing the harmfulness of mosquitoes to humans, without significantly affecting their role in the ecological balance between species, and the quality of the natural environment.
According to experiments already conducted, the laser does not need to kill these insects: it is enough to lightly burn the wings, enough to make them no longer functional. The mosquito falls to the ground and dries out.
I have no doubt that with the progress of artificial intelligence, and the falling price of high-performance hardware, this idea will very soon be effective, and it will be a very significant advance in the fight against the health scourges carried by mosquitoes.
Exactly! Need only “burn the wings” , in this paper i burn mosquito…
i want reduce the power of laser, make him safety for people and dangerous only for wing, need experience. Good research.
A 1W laser will blind innocent people nearby. The eyesafe limit is <5mW, so 1W is clearly dangerous.
No doubt someone will claim that the laser is moving or intermittent and so there won't be enough energy to burn someone's retina. That may (or may not) be true when the device is working correctly, but what's the worst case failure mechanism?
I don't really care if the designer blinds themself. I do care if they blind someone else.
Best way is to operate the laser in a box, where a co2 trap lures the mosquittos in. That also solves the problem of the best background to recognize the mosquittos.
Many ways to kill a mosquito, no need to use a laser. negative pressure into a mesh bag and let them die there is more than adequate.
i care about designer eyes too
There are eye-safe lasers operating in the 1.5 or 2um bands…these are strongly absorbed by the cornea which apart from needing a lot more energy to be damaged is fixable with today’s medicine, unlike the retina.
One option might be to use an event camera as it would be fast enough to detect the wing beat, thus making detection robust and very specific, so you wouldn’t kill/blind anything non-mosquito. It might even be robust enough to deploy on a drone, where the moving background would make detection of a tiny mosquito using cameras very hard otherwise. The issue at the moment is that event cameras aren’t really sensitive enough at the moment to detect the tiny movement of a mosquito’s flapping wings.
And of course you have to make sure no humans or other animals with eyes you don’t want to blind are nearby.
“And of course you have to make sure no humans or other animals with eyes you don’t want to blind are nearby.”
Like sharks, for example.
Laser safety goggles for sharks are plentiful on the dark markets. What you think those evil geniuses would want all the sharks in their moat blinding each other???
Thank you so much, information about the drone is very useful.
63 / 5000
Результаты перевода
Sounds familiar from 10 years ago.
There was a ted talk:
https://www.ted.com/talks/nathan_myhrvold_could_this_laser_zap_malaria/up-next?language=en#t-820389
I didn’t find papers about this device, only video demonstrations
No friggin lasers, but check out the Eave Tube: https://www.in2care.org/eave-tubes-work/
Thak you
it might be better for mosquitoes. But this device is universal, the mosquito is the hardest target, it is small and fast. But if we can destroy it, then the device will be able to destroy any pest in the field or weed
Or maybe use two moving laser beams, one low-power sweeping the airspace to be defended very fast or creating a “curtain” , just for target illumination and detection… and then aim the “kill laser” on illuminated small objects having a mosquito signature passing the “curtain” laser.
i thoug about it, it is good idea, that can make device more safety
I thought about it, it is a good idea, that can make device more safety
How about two low power lasers both sweeping. On detection by either laser, both lasers then focus in on the same spot where their combined power kills the mosquito. That way no individual laser is a danger, the only ‘danger’ spot is where they are both focused. Also your sweep rate is doubled. Scale up to more lasers if that financially viable.
good idea,
implementation of this idea makes the device a little more expensive, but if it makes the system safer as a result, then it’s worth it
Synchronizing the phases of two or more lasers is very difficult and requires nanometer level precision. Wind or dust could throw you off. Maybe if a third laser + sensor could give you the phase differential you could pull this off. But even the then – 2 class 1 ( <5mw) lasers won't necessarily be able to punch a hole in the wing fast enough.
Yes, I understand the dangers of a laser to the eyes, which is why I did not start crowdfunding. Need to find the safety value of laser power that can burn the wings without damaging the eyes. This is a separate study.
Also, this device can be used to kill weeds and pests.
Here is a published paper in a reputable journal.
https://www.mdpi.com/2073-4395/10/10/1616
Thank you for all comments, here really competent specialists
Have you considered replacing the laser with a “projectile” made of a sticky sugar solution? The solution, when it hit the mosquito, would cause them to fall to the ground immediately and the sugar in the solution would cause predators (ants, say) to eat the mosquitos almost as soon as they hit the ground. Also a sticky sugar “bullet” would be practically harmless even if it hit a person right in the eye. The only major problem would be hit percentage vs. a laser.
You don’t need a laser. You need to focus light in a spot near the mosquito. That would be much safer. You just need the location in 3D space and a way to accurately aim and recalibrate as needed.
How about a vacuum? Hey, seriously we need some robust OpenCV insect trained models. My house is full of termites, bed bugs, mosquitos and moths. It would be good to wake up my child if there are bugs crawling on his face.
There is a TED talk that did the same thing about 5 years ago
yes, of course, but in this case – low cost
To be fair, mosquitos don’t kill anyone deliberately. If we could cure them of malaria (and all the other diseases they pick up) then we could all live together happily. If a little itchily.
Make a small drone that can neutralize them upon contact. Oh, I think the creator has made mosquito hawks and dragonflies to do that.
yes, the final version should be installed in a drone
I think the answer is to setup a grid where the laser via mirrors creates a grid pattern where is any small flying bug be it mosquito or fly traverses the grid it will be killed. Could be sensors to detect the insect and instantly activates the laser.
Does not need to be super big say 3 x 3 ft. Then do something to attract mosquitos to fly into it such as a heat/carbon dioxide emitter or for flies something that emits a smell to attract them. I would make it with a really powerful co2 laser designed for laser cutting, 40 watt should do it. Sensor senses bug, laser activates and poof little pieces and smoke in the wind.
Make it portable but power it via 120v with an extension cord and place it outside. Make it so the laser does not escape the grid so it is relatively safe. Take it outside when the insects or active, turn it on and watch the fun.