Hi,
However, I don't recall a proper "specification" in this thread, so far. What type of birds*, e.g. size and colour, etc.? At what range do you wish to detect them and with what "field of view", i.e. the required acceptance (or rejection) angles of detection from the sensor?
There are various possible detection methods such as: PIR, Doppler "Radar", Doppler Ultrasonics or Visible (simple "image" recognition), etc.. If the birds have high optical contrast (e.g. against a light sky) then an optical method might be the most successful. Otherwise, I would consider Ultrasonic Doppler, using a reflector (possibly something like a small saucepan lid, but that can be calculated later) to define the detection beam width and range.
If my "back of envelope" calculations are correct, Doppler Ultrasonics and Radar are quite similar, but I think a bird may reflect more sound than electromagnetic energy back to a sensor (and IMHO 40 kHz sound is probably more "understandable" than 10.5 GHz electromagnetic waves). In both cases the wavelength is about 1 inch: 10.5 GHz at the speed of light (300,000 km/sec) gives a wavelength of 2.8 cms, 40 kHz at the speed of sound (~1,000 m/sec) gives 2.5 cms. Thus an object moving at 1 metre/sec (say 2mph, or slow walking pace) will create a doppler shift (frequency) of about 75 Hz (2 x 37.5, because both the "transmit" and "receive" distances are being shortened or extended).
So, if you can identify the speed of a typical bird (and/or its wings), it may be possible to devise a suitable (bandpass) audio filter (or maybe more than one) to allow the PICaxe to discrimate between birds and other (interference) moving objects. Not easy, but an interesting challenge. And even if it fails, the "bleeps" from an ultrasonic transducer might scare the birds away.
*EDIT: Ah, looking back I see it's pidgeons. Wood or Common?
Cheers, Alan.
Perhaps, because what you appear to be attempting is probably quite (or even very) "difficult".You are way above me.
However, I don't recall a proper "specification" in this thread, so far. What type of birds*, e.g. size and colour, etc.? At what range do you wish to detect them and with what "field of view", i.e. the required acceptance (or rejection) angles of detection from the sensor?
There are various possible detection methods such as: PIR, Doppler "Radar", Doppler Ultrasonics or Visible (simple "image" recognition), etc.. If the birds have high optical contrast (e.g. against a light sky) then an optical method might be the most successful. Otherwise, I would consider Ultrasonic Doppler, using a reflector (possibly something like a small saucepan lid, but that can be calculated later) to define the detection beam width and range.
If my "back of envelope" calculations are correct, Doppler Ultrasonics and Radar are quite similar, but I think a bird may reflect more sound than electromagnetic energy back to a sensor (and IMHO 40 kHz sound is probably more "understandable" than 10.5 GHz electromagnetic waves). In both cases the wavelength is about 1 inch: 10.5 GHz at the speed of light (300,000 km/sec) gives a wavelength of 2.8 cms, 40 kHz at the speed of sound (~1,000 m/sec) gives 2.5 cms. Thus an object moving at 1 metre/sec (say 2mph, or slow walking pace) will create a doppler shift (frequency) of about 75 Hz (2 x 37.5, because both the "transmit" and "receive" distances are being shortened or extended).
So, if you can identify the speed of a typical bird (and/or its wings), it may be possible to devise a suitable (bandpass) audio filter (or maybe more than one) to allow the PICaxe to discrimate between birds and other (interference) moving objects. Not easy, but an interesting challenge. And even if it fails, the "bleeps" from an ultrasonic transducer might scare the birds away.
*EDIT: Ah, looking back I see it's pidgeons. Wood or Common?
Cheers, Alan.
Last edited: