Room occupancy detector?

B

boriz

Senior Member
It’s easy enough to detect movement (Ultrasonic Doppler, PIR etc.) but how can you reliably sense whether a room is occupied or not? EG: Light’s come on when room is occupied and remain on until room becomes unoccupied, even if the occupant stays still for extended periods, like reading or whatever.

I can only think of one solution and I don’t like it. A beam break system with lot’s of little mirrors so that the beam bounces all around the room.

Any ideas?
 
P

papaof2

Senior Member
1. Motion detector(s) to detect initial entry.
2. Thermal sensors to detect warm body in specific locations in the room.
3. Pressure switches under the cushions or strain gauges on the furniture legs to detect weight of a body at rest in the room.
4. Beam break to detect movement into/out of the room (count number of entry/exit movements so a box left on a chair will not be seen as someone seated there). Count can be tricked if a child is carried into the room and walks out of it or the reverse.

There is probably some combination of the above that will work *most* of the time. I used motion sensors to control lights in the family room; worked fine when I watched TV but not for my wife - seems I move more when seated than my wife does...

John
 
J

Jeremy Leach

Senior Member
This isn't a good answer, but it amused me to think of it .... if you had a digital camera with a really wide-angle lens, then the digitised photo might change it's 'signature' if anything changed in the room. Ok, so anything changing would be detected, and exactly how to get this not to trigger because of light level changes etc is another problem - but it's an interesting idea ;)
 
B

Brietech

Senior Member
A beam-break (maybe just 2 across the door, a few inches apart, so you know which direction) that just tracks people in/out of the room is probably the easiest.
 
hippy

hippy

Ex-Staff (retired)
Jeremy's solution is probably best but not without some implementation problems. IR tends to be used in some schemes I've seen via Google.

Beam-break for counting people in and out works to an extent but isn't foolproof.

I'd probably go for PIR which works for most circumstances ( entry and when there's activity ) plus pressure mats under cushions where people may sit, assuming no one's falling asleep where they stand :)
 
L

LizzieB

Senior Member
Does the room have forced air ventilation? If so monitor the CO2 level of the return air.

I once worked at a secure facility that tried the in/out counting idea, we had fun with it (and the vendor) by often leaving through the windows.
 
K

kevrus

New Member
Working on the maintenance at a previous employer, a large office block was re-furbished. Part of the re-furbishment included PIRs in various offices to control the lighting. After a few months, they were all by-passed as people sat at an office desk didnt 'move' enough to keep the lights on. Various PIR locations were tried but it was purely a 'lack of movement' issue.

They could have all been sleeping though....
 
L

leftyretro

New Member
kevrus said:
Working on the maintenance at a previous employer, a large office block was re-furbished. Part of the re-furbishment included PIRs in various offices to control the lighting. After a few months, they were all by-passed as people sat at an office desk didnt 'move' enough to keep the lights on. Various PIR locations were tried but it was purely a 'lack of movement' issue.

They could have all been sleeping though....
Click to expand...
LOL, well I guess bypassing the PIR was a more cost effective solution then firing all the "low/slow movers" :p

Lefty
 
B

boriz

Senior Member
"I once worked at a secure facility that tried the in/out counting idea, we had fun with it (and the vendor) by often leaving through the windows."

LMAO
 
K

kevrus

New Member
Boriz, thats excellent, you could have an 'empty' building with hundreds (millions?) of staff present, or a 'fully occupied' building thats empty....
 
B

boriz

Senior Member
Taking Jeremy’s webcam idea and running with it…

Cam takes snapshot of room every second. It compares each new frame to the previous. If there is a difference, it assumes someone has entered the room and stores a copy of the previous, unoccupied frame. Then it uses that frame as the comparison frame for all further samples until it gets a match, meaning the room is no longer occupied. Then it goes back to comparing successive frames.

This would work, but only in a room with no windows and constant lighting. So some processing is required…

All the pixels are averaged, giving an average brightness for that frame. The average is subtracted from each pixel before storing it, so the resultant frame is a ‘deviation from the average’ image. Changes in ambient lighting, such as a cloud passing the sun, will only have a small effect on the deviation image. The system then just ignores changes that are smaller than this ‘ambient effect’.

Could work?

I’d still prefer a simpler solution though.
 
J

Jeremy Leach

Senior Member
Yep, something like that Boriz but I think it might be difficult to get working well in practice. For instance you bring the vacuum in the room then leave the room, but the software sees the room different because the vacuum is still there ...

Or you leave a cup in the room - the software needs to distinguish between big and small objects.
Or a door is left open and the room looks different.
Or the TV is on and the changing pictures make the room look different.

etc etc. That's why I said it wasn't a good answer ;) It's got potential in a very controlled environment ...but I don't think in a normal house.
 
Last edited:
You must log in or register to reply here.
Top