IR sensing to detect insects 2

[sean001]

I am interested in detecting insect movement with an IR emitter and detector. I have no experience with designing circuitry and need a good resource that will help me understand what I need in order to make this happen. Any assistance will be appreciated.

 

[itsmoked]

Wow that is a tall order as they will dissipate their heat so fast as to appear to be the same temperature as their environment.

If I were you I would start by asking IR imaging companies about if their product X can do this.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[sean001]

Sorry for the confusion. I'm not trying to detect the insect body heat. I want to detect when an insect crosses an IR beam.

 

[IRstuff]

Is this for school?

TTFN

 

[sean001]

This isn't a school project. The device will be used to study insect behavior for genetics research. I need to determine if the insects are moving past a specific point. I want to sense this movement with an IR beam.

 

[logbook]

You might try a slottted opto-switch. This will do the job for you provided the gap is reasonable and not svereal centimetres.

 

[nbucska]

Sean:
Please read my FAQ and give us some DATA.
The task is trivial but you have to define it with NUMBERS


Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[sean001]

Thanks for the reply, Nick. Here it goes. The insect is Drosophila Melanogaster which is quite small. The fly is confined in a 5mm OD (3.18mm ID) glass tube. I need to detect the fly as it passes a specific point inside the tube. Each time the fly passes the point, a microcontroller will record an activity count. The movement will need to be detected in both light and dark situations. Absolute darkness is required in dark stuations; not kind of dark or dark enough. Eventually I will want to monitor up to 50 flies at the same time, each in their own tubes. Right now I would be happy to monitor just one. I hope this information better explains what I'm trying to do. Thanks again.

 

[nbucska]

The sensor is easy .
Do you know if the D. is blind for IR ?
May the range of motion limited in vertical direction?
How much positional accuracy do you need?
Can you build the circuit if you get the schematics?
How intensive is the "light situation"? May the
light be fluorescent tube ?( No IR !)









Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[Comcokid]

From what you've stated, you should be able to do this with a photo-interrupter type switch. These devices have an emitter, and detector in one package with a slot between them. By selecting the right component, you may be able to drop the glass tube in the slot. These slots are made for cogged wheels or sliding tabs for mechanical control (your computer printer or photocopier uses them).

A quick check of some stocking parts distrubitors uncovered devices like (example): Fairchild Semiconductor H22A4

http://www.fairchildsemi.com/pf/H2/H22A4.html

I'm not saying this is the one, just an example.

Check part distributors like

http://www.digikey.com,

http://www.mouser,

http://www.newark.com,

http://www.jameco.com.

Use search terms like "opto switch" "Photo interrupter", etc. These sites will have links back to manufacturers data sheets. Then check the manufacturers web sites for applications notes which will give example circuits.

 

[sreid]

Another possibility from Onron.

https://omrwsc.am.omron.com/webapp/...1&productId=115061&langId=-1&categoryId=15720

 

[OperaHouse]

I'd consider one emitter and two small detectors side by side. You never know, the body parts may transmit light fairly well or not block enough. A differential detector may be a lot more predictable.

 

[VEBill]

It's likely that such wee little bugs won't properly block the IR with a plain-Jane off-the-shelf photo-interrupter switch tossed around the glass tube. The bugs are almost certainly too small (2mm ?) to block the beam, and also possibly too transparent.

Those sort of switches are probably designed for something large and solid to block the IR. I wouldn't even assume that they could be reliably adjusted to work (adjusted perhaps, but reliably maybe not).

Also, the glass tube might bring some optical complications. If the IR beam were very narrow, then the tube simplifies into two walls (optically simple). I think most such sensors have larger beam widths.

I don't think it is necessarily as simple a problem as simply using an off-the-shelf photo-interrupter switch. It might be, but perhaps not. Small beam width would help.

If the problem was lab rats in a 10-cm tube, then it really would be easy. But tiny bugs in a 3.18mm ID tube brings a few complications (or at least design risks).

And 'nbucska' had a VERY good (biology) question about the bugs' opinion about 'absolute darkness' versus being blasted with near IR. Camcorders pick-up near IR so it isn't a stretch to wonder about the wee little bugs needing sunglasses.

If you do use optical sensing, maybe it needs to go end-to-end through the length of the tube (with entrance and exit ports for the bugs). Might require custom circuits to be reliable. I'm assuming the bugs have better IR blockage lengthwise. Plus the signal would be longer duration.

Alternate concepts:

Maybe the bugs could walk across a microphone element ('scratch scratch scratch'). Would require a sensitive element, a low noise preamplifer, and a quiet environment. Could use multiple elements and timing/logic to reject ambient noise and determine direction.

Maybe a (very very) sensitive balance (using mass). Tricky.

Maybe ultrasonic Doppler (affect the bugs?).

Maybe vision systems (IR again).

 

[itsmoked]

Just make each bug carry a pass key.


Actually I agree with VE1BLL on the run-of-the-mill interrupter switches being iffy.

1) I would design an analog system that essentially looks for a 'dip' in the light as this would allow a fly to walk thru anywhere, upside down, etc and a dip would still occur.

2) I would put two sensors right next to each other and look for the one-two punch,(dip-dip),to confirm Mr. Bugs passage.

3) Consider using reflective technology so the light shines against the bug body and the reflected light confirms bug passage.

4) If any kind of daylight can get anywhere near this system then modulation is absolutely required if you want reliability. With the reflective system you just look on the receiver for the modulation signal to appear,(=bug).

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[MikeHalloran]

You want a photo- interrupter that does NOT have a Schmitt trigger in its output, so you can detect small changes in photocurrent to deal with small targets that are probably not opaque to IR. The beam of a typical interrupter is about 1mm diameter; the tube will distort it a little, probably in a way that works in your favor.

You'll probably also need to adjust the source current. I'd start at a couple of milliamperes and work up slowly from there.

Daylight and incandescent lights will blind the sensor, fluorescent lights won't. I like the idea of modulating the source and just looking for that frequency on the receiver. You don't need a fancy tuned circuit in the receiver to implement synchrnonous detection, since you've got the modulation signal a few cm away already.

Find a friend who can explain 'synchronous detection' in English and you've got your circuit designer.





Mike Halloran
Pembroke Pines, FL, USA

 

[cbarn24050]

It's not that hard, an interrupter with transistor output ac coupled to an opamp then onto AD input of the micro through low pass CR filter. Drive the led with PWM, self adjusting the duty to give you a fixed level at AD input. Insect will chage the level allowing you to count it.

 

[VEBill]

OperaHouse's differential detector is a good suggestion.

 

[sean001]

Nick,

1. Do you know if the D. is blind for IR ?

IR is used with the flies in similar applications and there is no report that it has any effect on them.

2. May the range of motion limited in vertical direction?

The tubes will be horizontal. They are about 70mm in length. The fly can walk around freely inside the tube. There is no room for it to fly.

3. How much positional accuracy do you need?

Not much is needed. The idea is to see if the fly is moving. It doesn't need to be monitored in an exact location. If it moves close to the detection zone and then turns around but is detected, that is good.

4. Can you build the circuit if you get the schematics?

Yes. I should be able to build it with schematics.

5. How intensive is the "light situation"? May the
light be fluorescent tube ?( No IR !)

I'm told the light cycle does use fluorescent tubes. I'm also told that similar devices use IR detection in the presence of fluorescent light. Based on what you and a couple of other replies mentioned, this should not be the case! Am I getting incorrect information from the fly people? The flies are in a temperature regulated incubator so I'm sure the fluorescent lights are correct. Is there a way to use IR detection in fluorescent light?

Thanks to everyone for the advice.

 

[nbucska]

The modulation is certainly good advice -- it simplifies
the detection but complicates the circuitry.

If you want to detect the motion -- any motion --
perhaps video camera would be better.

Use -- perhaps vacuum-formed -- mini-Petri- dishes
arranged in an 8 x 8 matrix or it may even be one piece.

the video signal is digitized and iinterface to a PC. The
PC stores the picture in an array form, compares it with
the previous picture pixel -to pixel, and if there
is/isn't any difference, determines its position by cell
number.

You need a cheapest camera and inteface card, off the
shelf, no circuit to build. The software is trivially
simple.



Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[VEBill]

But he want the system to also work in the absolute dark. So the camera would need to be IR and he'd use IR illumination (assuming the bugs are blind to IR).


I'm amazed that nobody has wandered by and suggested RFID tags... ;-)