Forum

Movement detection

hi all,

just a question
what’s the best/cheapper way to detect an object (eg:people) in a given field/aera?
assuming the light in this aera can change(eg:stage)
and the detection need to be precise to trig events
each time an object is detected:

1/ distance sensor (not so cheap)
2/ Volumetric movement detector (cheap security stuff)
but need arduino hardware…to get the value back in vvvv?
3/(IR) cam (light condition can be a problem)
4/ ?

any advice welcome
thanks

Hi circuitb, for my own i have made the following choice, wich is very good working and is not posing any troubles with changing environnement ( lights etc):
-Normal camera, wich you have removed infra red filter (webcam / dv, old vhs camescope, anything)
-a filter to cut off everything else than IR ( photo film developped but not exposed)
-IR leds 880nm -> a solution very cheap and very reliable. there is also IR spots.

Precise definition: i would go on a 60fps cam, and no dv ( interlacing troubles, especially for fiducials)

;-)

-If you are doing it “on stage” there’ll be plenty of light bulbs producing a lot of IR-lights (light bulbs are giving 10-15% visible light only!, rest is IR and heat-IR).
So, choosing IR for movement detection will cause troubles as light changes will easily interpreted as movement then.
Usually you use IR-cams to split visible and unvisible-IR-light for cams and people, assuming you are using separate lighting as well. Which you arent in this case - the stage.
-Same with “Volumetric movement detector” or PIR-sensors. They are looking for IR-heat changes and giving you an event then. Advantage is that they adjust to new “heat” situations quite fast. But tracking precisly inside a spotlight (by hot light bulbs) will hardly giving you an event. Ever tied to use this PIR in summer? You have to be hotter than your surroundings. So, very likely to be unsuitable.

How is the situation exactly? Can you throw IR-light from side or behind or background? How’s the floor? Reflectable of not?

thanks for replies:

the situation take place on a standard stage (8 x 4 meters)
i 've to check the floor material…
the lights are also standard (front/side/back)
i’ld like to define 2 aeras on both side of the stage
then i think it’s possible to have 2 cam/sensor on both side
with their IR spots…
what i understand is : in any case the surrounding
heat/lighting will cause unwanted events (i can’t filter the whole stage lighting system)
so i’ve to check with the light guy to have a static light
situation and make the sensitivity calibration during the sound check

for the PIR you mean a temperature above 38/40°C on stage could be a problem?

perhaps an ultrasound sensor…like a sonar could be
solution :)

for the PIR you mean a temperature above 38/40°C on stage could be a problem?
-Very likely, they are looking for heat changes to surroundings. So, if your surroundings are 38-40°C your actor should have fever to be recognized by PIR. In theory, in practise is best to just try it with a cheap PIR-lamp (the one you can attach outside your house).

-With cams you can shoot into the direction you have no light fixtures: from back to front. Put cams into an angle that they are not getting any floor (unwanted shadows/lihgts), but the dark audience only. So, you have only lightened actors or dark background - but no unlightened actors!
Thats a way it could be done, depending on stage and light design.
May be you have to cope some regions in this video image by hand (or black quads), where there are small lights for emergency exits. But it shouldnt be huge areas.

Ultrasound is best choice, in fact, but video is the easiest in practise ;)

ir thermography cameras are also quite usable in this context. they usually have an composite video output and are therefore very easy to connect to vvvv

see e.g. http://www.flirthermography.com

thx frank

oschatz :the FLIR looks good
i’ve seen they use it more and more to check thermal insulation
but at this time it’s still very expansive…5750€ for the cam…