Infrared sensor array to detect mosquito entry

[Vizier87]

Hi everyone.

I've been attempting to construct a simple mosquito entry sensor array using infrared photodiode pairs. They kinda worked, but the problem is the line of detection is too fine.

Ideally, the opening where the mosquitoes enter/exit is about 10cm x 10cm. That's where the sensors are installed.

I just wonder before I engineer the heck out of the setup, are there readily available sensor arrays I can use which work?

And yes, I have used cameras before. They work, but sometimes those bugs zip past through too fast and it doesn't detect it .

Appreciate the input.

Vizier87

 

[danadak]

I wonder if wing beat frequency could be basis for detection.....

Currently image recognition would work, buts escalates costs/complexity.

 

[Tony Stewart]

With a 200 mm X+Y linear aperture, you might see a < 0.1% shift in transmitted light down in your noise level and need a very low reflected background for reflected light. An audio mic and bandpass filter would be needed with two back-to-back mics for the direction of motion and confirming they enter or exit to hear the bzzz like when you are sleeping.

 

[Vizier87]

I wonder if wing beat frequency could be basis for detection.....

Currently image recognition would work, buts escalates costs/complexity.

Yes, ultrasonic types have been around but to my knowledge they're not much of an option since they need to be pretty well filtered. Plus what I need is pretty much an "entrance" sensor.

 

[Vizier87]

With a 200 mm X+Y linear aperture, you might see a < 0.1% shift in transmitted light down in your noise level and need a very low reflected background for reflected light. An audio mic and bandpass filter would be needed with two back-to-back mics for the direction of motion and confirming they enter or exit to hear the bzzz like when you are sleeping.

What would you recommend? I've never used audio based types but pretty willing to try it out if it actually can be better

 

[Tony Stewart]

Desktop mics use this. https://en.wikipedia.org/wiki/Electret_microphone < read
The noise cancelling feature uses a hole in the rear for cancellation but may need balancing.
Pull-up with 5.6k resistor from a well-filtered 5V supply.

Ebay , Digikey, mouser etc

 

[Tony Stewart]

If you plan to use this in a noisy environment, it may be hard to detect the difference in 2 mics accurately to detect a mossy by having them on opposite sides. But if identical sound pressures from ambient, this might be OK if the middle is partitioned. Then a differential amplifier to a half-wave detector can cancel distant sounds hopefully more than the nearby wing noise. The spectrum is harmonic in nature but dominant near 350 Hz but filtering may not improve SNR much.

 

[danadak]

Yes, ultrasonic types have been around but to my knowledge they're not much of an option since they need to be pretty well filtered. Plus what I need is pretty much an "entrance" sensor.

https://www.researchgate.net/figure/Wingbeat-frequencies-for-29-mosquito-species-Bars-mean-wingbeat-frequency-are_fig1_351489590

 

[Tony Stewart]

The above data was taken with dual IR sensors to detect wing frequency and direction of travel in a much smaller aperture of 1.5 cm sq. using alternating UV LEDs to attract them.

The detector was high speed regulated to 1.5V by modulating the emitter voltage if the error exceeded 50 mV. A Mosquito entry was detected by a 100 mV drop for 200 ms capture and analyzing the data for fundamental. I would expect 1f, 2f , 3f and 4f modulation from combinations of a single wingbeat or a bidirectional wingbeat attenuation and then mixed with dual wingbeats.

I would expect the sequencing of these events with 2 detectors of returning to non-blocking to validate the direction. This quasi AGC could be done digitally or by an integrator feedback from the Rx PD much slower than the mosquito transit time.

Event counting requires a small aperture to be accurate yet something to motivate them.

I have designed optical detectors that can detect the wire of a 1/4W resistor in a 1m path line of sight using recessed emitters and a recessed IRDA Rx.

What is your purpose?

 

[Tony Stewart]

if your initial effort worked but too fine(?) then you just need an array of detectors from a single source or a linear array of narrow sources

 

[Tony Stewart]

Acoustic with dual electret mics will perform better far field noise cancellation than with higher sensitivity for near field than one . BTW They have a back cloth to attenuate cancellation effects but some are better than others.