ideas for modelling false alarms detection from animals for a P.I.R sensor alarm?

[Jasmine]

Hello,

I want to build a P.I.R sensor alarm which can distinguish between a human body and small animals based on their differences in mass(hence less IR radiation from the animal). I am thinking of designing a separate circuit with an IR source to model different animals and to test it with the P.I.R sensor to get the optimum performance for animal detection. Can anybody provide me with ideas about how to do this in order to obtain accurate results...?

[Charles Yarnold]

Hi Jasmine,

What's your end goal? If you're not doing this for the fun of building it from scratch you may be able to use an off the shelf sensor. You can get hold of pir sensors that will ignore animals smaller than human size, these are used in alarm systems where you don't want to have to lock your pets into one room of the house.

If you use one of these and a standard PIR together you can check if a trigger is human or animal, i.e.

both alarms go off = its a human

the normal pir goes off but not the other = animal

Best,

Sol

On 2 October 2013 19:59, Jasmine <mehja...@gmail.com> wrote:

Hello,

I want to build a P.I.R sensor alarm which can distinguish between a human body and small animals based on their differences in mass(hence less IR radiation from the animal). I am thinking of designing a separate circuit with an IR source to model different animals and to test it with the P.I.R sensor to get the optimum performance for animal detection. Can anybody provide me with ideas about how to do this in order to obtain accurate results...?

--
You received this message because you are subscribed to the Google Groups "London Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to london-hack-sp...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Nigel Worsley]

> I want to build a P.I.R sensor alarm which can distinguish between a human
> body and small animals based on their differences in mass(hence less IR
> radiation from the animal).

The problem with that is working out how to tell the difference
between a large object at a distance and a smaller one closer to the
detector. Or a really small object really close to the detector, such
as a spider crawling over the lens - this is a common cause of false
alarms as they like to hide behind the detector and then come out at
night!

> I am thinking of designing a separate circuit with an IR source to model
> different animals and to test it

What do you intend to use for the IR source? Forget LEDs, the required
wavelength is about 10 times that of a typical IR LED.

Commercial PIR sensors are tested by pointing them at a heated wall
and then moving a room temperature target in front of the wall. The
fact that the target is colder than ambient rather than hotter makes
no difference, as the detector elements consist of two sensing areas
side by side, and the difference between the two is used as the input
to the processing algorithm. The detector cannot tell the difference
between a hot object moving into the view of one half of the detector
and a cold one moving into the view of the other, the view usually
consists of an array of 'beams' due to the multiple fresnel lenses of
the detector.

It has been common practice for decent PIR detectors (ie. not the
typical crap used for exterior lights) to use a microprocessor for
signal analysis for over 20 years.

Nigle

[Adrian Godwin]

I'm looking at some thermal sensors with 16 elements. I'm not sure if they'll be helpful as I haven't analysed their response yet.

On the more advanced types that Nigel refers to, what signals are processed ? Is it still based on the 2-part differential sensor or do they use something with more resolution ?

[Jasmine]

Hello Solexious,

It is a school project actually and I have to build it from scratch. The end goal is to build a P.I.R sensor alarm using analog and digital techniques to be used outdoor. But I must find a way to minimise false alarms caused by pets and demonstrate it.

Thanks,

Jas

[Jasmine]

Hello Nigle,

Actually I am unsure about the IR source. I was going to use an IR LED. I thought the sensor response was independent of wavelength but rather dependent on the amount of incident radiation. That's why I posted here to get some ideas about this part of my design. I only have school laboratory components to use.

Jas

[Toby Catlin]

I don't have much experience with openCV and the like, but I know others on the board do. Would it be possible to use a night vision camera (IR + IR lights) to recognise movement and size of creature?

--

[Nick Johnson]

As others have mentioned, an IR LED isn't going to work. A lightbulb might, though, as might an electric element with a thermostat set to body temperature.

-Nick

--

[Mehjabeen]

The P.I.R sensor circuit is fairly simple. I dn't know how to relate it to a camera with recognition parameters.

--
You received this message because you are subscribed to a topic in the Google Groups "London Hackspace" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/london-hack-space/2teOYXdhja8/unsubscribe.
To unsubscribe from this group and all of its topics, send an email to london-hack-sp...@googlegroups.com.