Open Source Mosquito Locator

I’m looking for help with a project that will provide a way to locate mosquitos. Yes I am serious.

A Mosquito on an arm

Remember this is not the army. We don’t “locate and destroy”. We’re following the Unix  philosophy of doing one job and doing it well. This job is to “Locate“.

It should be simple.

It should be cheap to make and cheap to run.

There is no requirement to identify species, gender, age, colour, etc. Let “if it’s acting like a mosquito then it’s a mosquito” be your mantra.

Let’s be clear, this is just to locate them not to terminate them with lasers, automatic weapons, flame throwers, etc. If you want to do that later then be careful. For now the task is to locate them and signal where they are. When they move, locate them and signal where they are.

I have absolutely nothing to contribute to this project, but it needs to be done.

If you have any ideas on how you would go about doing this, then I’m very interested in hearing from you.

 

This entry was posted in General. Bookmark the permalink.

21 Responses to Open Source Mosquito Locator

  1. TuxSax says:

    Sounds like a great idea, I’m not sure what are you planning with the info once you “locate” the mosquito. I guess you are only planning to supply the information and let the user do whatever they want with it, use the laser, captivate the mosquito, splat it, etc…
    Here’s a question about your project.
    What is the biggest area you’re planning to locate the mosquito in? A house? A Park?
    Do you want to locate them one by one, a few, hundreds?
    I’m asking because this fact will dramatically affect choosing from the possible ways to do it.
    I’m not an engineer but from my general knowledge I can only think about two ways technology provides you, which I’m not sure if it’s feasible for this use at all.
    1. Ultrasonic, similar to how you see a baby inside his mother’s belly, or how bats “see” with their ears, analyzing the way the echo of the sound waves comes back.
    2. Based on mosquito sound itself, but you’ll need a very sensitive microphone and a very sophisticated way to filter all other irrelevant sounds captured from the environment, which in terms of very high definition microphone, could be very very noisy, so you need to find a way to pinpoint the very single noise frequency you’re interested in. You could also be able to determine if it’s a male or a female mosquito as their buzz sound is different.
    3. Thermic camera? If you manage to settle on the heat footprint of a mosquito (if there’s such a thing…) You could detect it wherever you want, including through obstacles.
    Again, I think the range and topology may be a big challenge here, so if you’re trying to find a mosquito in a house, you’ll have to install a sensor on every room, and make sure the sensor can detect a mosquito behind a cupboard, for example, which I don’t believe it will be easy.
    I hope my inputs can help you develop the idea

    • ken_fallon says:

      I guess you are only planning to supply the information and let the user do whatever they want with it, use the laser, captivate the mosquito, splat it, etc…

      Exactly.

      What is the biggest area you’re planning to locate the mosquito in? A house? A Park?

      Let’s assume the easiest situation for now, so assume a baby’s bedroom.

      Do you want to locate them one by one, a few, hundreds?

      Just one.
      A simple “There (was/was not) a mosquito detected since I as activated” would be a good start.

      You could also be able to determine if it’s a male or a female mosquito as their buzz sound is different.

      Yes I know and only the females bite but don’t let the fact make the project more complex.
      If identifying a mosquito by sex makes it more complicated then ignore the fact and identify all mosquito’s.

      you’ll have to install a sensor on every room

      If it’s cheap enough then fine, and it can also me moved.

      and make sure the sensor can detect a mosquito behind a cupboard

      NO !!! Keep it Simple. If it stays behind the cupboard then fine by me. When it comes out then tell me about it.

  2. TuxSax says:

    So I guess you’re talking about an invisible wide spacial “mosquito net” like dome, that is pulled on the baby’s room space to detect and alert about the presence of a mosquito?
    That could be great indeed.
    Another idea I thought of is to use some kind of movement or volume detector, similar to what house and cars alarms use, perhaps a laser curtain?
    But in this case, the software that uses the sensors will have to know to set a room footprint for things we know are there and we don’t care about, the baby moves, there are toys and furniture around, perhaps a fan is working, the curtains moved by the fan breeze, etc.
    Once the baseline is set, it may alert on any event that goes over the threshold that was set up.
    Again, I only have ideas, but I don’t have a clue about HOW can that be done…
    I think you should try starting by researching on the many available types of sensors you can use to detect such a small insect, perhaps a combination of two of them will give better and more accurate results, and not alert you about some other same size flying harmless insects.

  3. J says:

    Any progress with this? I was thinking of setting up triangulation with three microphones(or better 4). I came across https://instruct1.cit.cornell.edu/courses/ee476/FinalProjects/s2007/sp369_bb226/sp369_bb226/index.htm some years ago, but they had numerical problems(something about their jacobian being ill conditioned) making the result inaccurate. Maybe I could build something with an arduino or rasberry pi. I think there are ways to filter background noise(if all microphones hear it well).

  4. marc t says:

    http://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=1055&context=abe_eng_pubs

    Detecting Insect Flight Sounds in the Field: Implications for Acoustical Counting of Mosquitoes

  5. ken_fallon says:

    First of all thank you for replying. I am still looking into this project but progress has been unfortunately very slow.

    Your idea of using drones seems like a great solution and thanks for thinking about the problem. There are a few issues that I would like you to think about and see if you agree with my logic. Firstly the cost of drones is quite expensive and we should be aiming to make this as cheap, accessible, and hackable as possible. The cheapest drone I could find was €129 and I’m not sure how hackable it would be.

    The next issue is that something as small as a mosquito will only register as a few pixels on a cheap camera, and on a higher resolution camera the cost increases as does the cost of the computer required to process it. You could definitely use imaging on a well lit white wall with a high resolution camera. To process this you might get away with a RasperryPi 2 with the Camera Module and this will easily cost €100 or more. In impoverished areas one cannot assume that they would have the money to be able to run a light all night, and it would also interfere with peoples sleep pattern. You are going to have to add to the cost by having the drone take an image from exactly the same position every time. If it’s out at all then the entire image will seem to have changed. While it probably is possible using very expensive laser guided drones, it probably would be cheaper to use a lot of professional fixed cameras. Which is to say very expensive. Furthermore the system would be off line while the drone charges.

    Even with fixed cameras the image would change in low light conditions as dust or other air borne particles move across the image. In a child’s room the image would change each time they shifted in their sleep. That said, I have not considered the option of imaging in the non visible bands where it may be easier to “see” the mosquito. I’m still not sure how easy it would be to be able to tell a mosquito from a house fly for instance. Anyway while this seems like I’m knocking your idea, I’m not. It’s very possible that a small micro drone that is light and mass producible could be developed that chases the mosquito. But for this I would like to get it down to reusing currently available mass produced cheap electronic components.

    Right now I want to get as much information on mosquitoes so that we can see how to identify them by sight, sound, vibrations, radiation, or whatever. Then we can see if there is a cheap sensor that can identify that, then we can see if we can do any calculations necessary in as small and affordable device as possible – preferably basic electronic components, if not a small arduino, if not a raspberryPi, and failing that offloading the computational work to “the cloud”.

  6. Alexandre Azzalini says:

    Hello,

    have you seen this site ?

    http://humbug.robots.ox.ac.uk/research/sensing/

    Kind regards,

    Alex

  7. ken_fallon says:

    Hi Alexandre,

    Thank you so much for pointing me towards this project. I got in touch and Recorded an interview with them for Hacker Public Radio.

    http://hackerpublicradio.org/eps.php?id=1894

    Ken.

  8. Alexandre Azzalini says:

    No worries !

  9. Maurizio says:

    Hello!
    Any update on this? I was starting to think about this exact issue my self, realizing that this is probably the hardest part of any electronic mosquito killing device. It seams a damn hard problem, I have no idea about where to start, just to be clear. It looks to me that microphones or cameras should have a pretty high resolution in order to spot mosquitos, let alone to triangulate their position.

  10. Andrea says:

    Dear Ken,
    last summer after several mosquito bites I was planning a revenge and going to spend some time in doing what you planned. However I stopped for lack of time.
    Are you still working on it?

    Andrea

  11. Borg says:

    I also to research for mosquito detection or shows by specific by image sensor.
    As we known that the mosquito is too small in image sensor, thus I want to show mosquito by display device or glass.
    maybe the specific IR spectrum to show or display by blood in mosquito/mosquito body/wing.

    I also want to co-work with your in future.
    Thanks

  12. Tyrone lee says:

    How about using a thermal array eg Heimann sensors?

  13. Nidhin says:

    Hai ken_fallon , did you complete your project?

  14. Santa is coming!

    ‘Laser Movable Mosquito Killer Robot’ (made in China)

    https://quillorcapture.com/2016/09/14/presenting-the-mosquito-killer-robot/

  15. Hey all. I landed here trying to find a way to locate mosquitoes too.
    Apart from mosquito repellents, I feel there is very little effort put into mosquito detection. Its all just sprays and creams. By the way am allergic to mosquito repellent spray.
    Right now am doing some bmotion detection work on my Linux box: as pointed out, it’s next to impossible to detect those pests using a camera cos of the noise. Also, sensitive microphones would pick up a lot of noise too. Am in Africa and those pests are all over the place. All I can add is that this needs to be done, somehow.

  16. Ehizojie Orhiere says:

    Hello
    http://m.caltech.edu/news/mosquitoes-use-smell-see-their-hosts-47338
    Check that link out.
    How about making the mosquitoes come to the device? They come to the device to feast and get trapped. Then they can be counted. The device may also crawl or fly around… If the mosquitoes get close enough then they can be seen by a regular camera. And detected by microphones.
    I really want this project to succeed…. I’m from Nigeria…. Apparently I’m never going to get used to the bites…. And I get malaria alot
    I’ll keep thinking up ideas…. Just decided to search online today and I found that link above and this one too
    Power to the drones!!! 😃

  17. Anonymous says:

    While the already cited Mosquito Fence isn’t open, I think there is an article in wich they describe how to do it with common parts of hardware for a price of less than a 100 dollars. (Wich I believe was the goal of the original team)…..and i had one of these disgusting insects interrupt me while writing this post….

  18. Peter Zelchenko says:

    We have them bad in Shanghai. If your sensors are too costly / insensitive, then the above advice of bringing the sample to the sensor — well, it uh makes sense. There are a number of things that attract mosquitoes to targets, so the attractor is probably simple and cheap. Getting the sample to settle or hover long enough near the target is thus easier.

  19. Blank says:

    With what sensors you should start out to detect them in room? Can you filter out all noise with good microphone and detect its wing beat? How about thermal or night vision cameras combined with wing beat detection? Or maybe there is some method to amplify one specific small area sound strength?

  20. Anonymous says:

    If there is some sort of ultrasonic mosquito signature, I would love a device that could help locate the source of mosquitoes in my neighborhood. Something directional so you could follow the swarms or take samples from multiple locations just to get a rough count could be useful.

    My city has installed a network of gunshot detection microphones and something similar for mass mosquito control seems like it could have plenty of benefits for everyone.

Leave a Reply

Your email address will not be published. Required fields are marked *