Heat Map, CO2 as presence


Hi Guys,

I was noticing this the other day.
It seems that temp sensors have a range too.
I tested this with a few different brands and with non smart sensors as well.

The temp inside your house will vary from spot to spot and is not even through the house no matter how good your insulation is this happens because of the laws of thermodynamics.

So I was thinking, if you know the range of your temp sensors, if they are sensitive enough and you have plenty enough you could draw a heat map of your house. This is theory.

What you guys think?

P.s @veeceeoh would you know if the Aqara or Xiaomi original sensors support readings beyond first decimal (××,×), for example xx,xxxx? Would be interesting to understand how sensitive they can report.
For this purpose the pressure measure on the Aqara would also be interesting to know.


I think you might get a "general" idea of warm and cool spots. However I don't think the results would be worth the effort.
I would contact my local power company and ask them to do an efficiency review which in my area includes an flir photo.


The idea is not related to if there is or not cool spots. But rather use heat and cold patterns to detect heating variations.

Let's say for example if I lit a candle if that would be detected. Just an academic exercise really.


I’m not sure I understand the goal of doing this. Can you elaborate further?


The theoretical idea is:
If we can detect the heat map and small temperature variations on it we can extrapolate when someone enter a room. Additionally we can potentially also track movement patterns, with enough to data we can than use that movement patterns to make the house automations more sensitive to flow patterns.


Is it common for temperature change in a room to be used as a proxy for presence detection?


Don't know... But it's not relevant.
The science behind supports it.
As an example similar approach used by the military on thermal Google's.


Hmm, I’m not sure how infrared goggles are relevant.

Humans radiate heat into their surroundings, but does that raise the room temperature around them in a way that could be reliably detected and used to indicate that a person is present?

That seems like a very different question.

And I guess what I was asking was, has anyone ever done this in any other contexts? If so, then the next question is, will this work with the temp sensors that are in most z-wave and zigbee devices?

If this has never been done before, it suggests to me it’s unlikely to work no matter what temperature sensor is used.


Agree that's not easy as the mathematics are not yet all understood, in particular how Havcs, wall density and structure, window type etc will play a role.
However with multiple sensors that are sensitive enough in theory should be possible to pinpoint a heat spike.

This was the paper that make me start thinking of it.


On a side note I'm already using Co2 levels to check if a room is being used or not.
It uses triggered window sensors as well. In fact is the only reason why I still have ST plugged in. Nobody has fully implemented Netatmo into HE.


Thanks for sharing that article, I will take a look.

I think this could be a more specific environmental change that would signal the presence of a human, or maybe a pet. So many other reasons why the temperature could change other than because of heat radiating off of a nearby person.

Edit: link doesn’t seem to work.


Edited, it should automatically download the PDF file


It is working for us (we have no pets).
We figured out (thanks to Netatmo saving a history of the readings) that as someone is in a room the Co2 levels raise @ average of 25 units per person until stabilize. Roughly half of that amount when is empty will decrease until it stabilizes.

The challenge is that Netatmo only uploads data every 10.. So not useful for instant presence detection


But the co2 levels wouldn’t necessarily rise instantly, would they?

I think that PIR sensors do the job better than most others, though they have their shortcomings too. But they use IR to detect the heat radiating off a person, just like those military goggles, so that makes them badass :sunglasses:.


Interesting discussion. I have a very open floor plan on the main parts of the house so could this be very a different exercise.

I would assume that the ability to trigger a temp reading based on motion activity could lead to interesting data too.


It's gradually, so it's possible to be used. And yes as soon you enter a room and brief into the air you are increasingly the c02 levels. The bigger the room the more you need to breath to affect the results. So it works better in smaller spaces.

Let me take a screenshot of some data just to show you
As you can see I enter the room @22.44 (I cross checked with the contact sensor on the door) the Co2 levels raise gradually during all the time I was in the room. Each dot is a 5m interval.

Edited Before someone asks :slight_smile:
The Normal values for PPM in my area is 350/400 PPM.
The reason why is so low is that's quite a cold evening tonight and I had left the window open all day so the room was more than properly ventilated.


How would you approach it? Motion sensors as far I know can't trigger temp readings...


That's exactly the same principle for some Co2 Sensors. This Co2 Sensors use IR light to measure C02 particles.

"Your sensor measures CO2 using an optical method: Smart Indoor Air Quality Monitor contains a light bulb and an infrared sensor. The light emitted by the light bulb is partially absorbed by the CO2 contained in the ambient air. The higher the CO2 level, the more the light is absorbed. The infrared sensor measures the quality of light received and deduces from it the level of ambient CO2.
This method works independently of the level of ambient light."


Sorry, Multi-sensors could do both.

I have zoned radiant heat in every area and with the open floor so something like this might not be fully possible but interesting none the less.


You just raised an interesting question.
Is the radiant walled or flooring?

Don't have any multi sensors so not real familiar how they report the data.
But for example the last version of the Aqara motion sensor will provide a lux value when motion is triggered.
If the multi sensor you refer to works similar yes, it raises interesting possibilities too.


Ceiling actually, very old Ranch style home, has pros and cons. Rest of the heat comes from a pellet stove in the basement. Heat from both kinda meets in the middle.

I have 3 Aeon Multi-Sensor 6s. Haven’t played with them much more than simple motion. Off the top of my head they do motion, temp, humidity, and lux.