morning

@ThePhoton perhaps you could say a light sensor

and you want to subtract ambient light from a source light

and the ambient light level can vary

the method i'm using is to sample the ambient light and then sample the source light and then subtract the ambient light from that

but the problem is if it is automated how to know when the source influence or other external influences could be incorrectly raising the ambient level when the ambient level is sampled

i think a possible path would be to have directional array and detect external sources that way and try to determine the background based on that information . . but i am not sure

morning

@jippie morning

@EwokNightmares hi

@jippie hi

how are your cats

?

@EwokNightmares eating

@EwokNightmares they sound like a fire engine when I take them on a field trip (to the vet)

@EwokNightmares Yes, that's a very common correction we do in my work. We call it "dark calibration". Although mostly we use it to account for leakage current through the photodetector, and we try to avoid ambient light altogether.

I was worried you were thinking of something like an RF detector and trying to detect a signal below SNR=1, which I think there are ways to do it but it is not necessarily easy.

@ThePhoton Ok, so then how could you do this correction "blindly"?

say it is a portable IR intensity thermometer or light intensity detector

@EwokNightmares In our setups, we turn off our light source and measure the background, then turn on our light source and do the measurement.

But my setup is optical fiber where ambient light is not really an issue, just detector leakage.

Ah

this is a portable detector that could be exposed to the light source or other light sources at any time it is expected to sample the background

@EwokNightmares Can you modulate the light source you want to measure and use a lock-in?

@ThePhoton cannot modulate other than to sample it

the source itself will be fixed

the problem is say someone is outside and takes the bakcground then steps in a low light room and now tries to measure the light intensity from a source in there, then the background subtraction will give a very high artificial

@EwokNightmares Fundamentally, you have to have some way to tell the instrument "This is the background" and "This is background plus signal".

error*

@EwokNightmares That's a kind of universal problem with any calibration procedure. If you calibrate, and then change the error sources, the calibration is no longer valid.

@ThePhoton yes this is basically a zeroing of a scale just a complicated zeroing

i was thinking if i had sensor array, i could determine the direction of intensity and if one direction were identified as the source, then that source's intensity as it attenuates to the other sensors in the array could be calculated with 1/r^2 and attenuation factors of the materials and then those computed values could be subtracted from all the other directions and averaged for an average background

@EwokNightmares That's one idea. I had thought of it with just two sensors. The problem is how to point the "background" sensor (or choose the array element to take as background) to accurately measure the same background light, without being corrupted by the signal light.

i guess that doesnt get me very far unless the source is a point source

hmm

I think the multiple-detectors strategy needs operator judgement to work well, and it sounds like you're trying to avoid that.

i wonder about a hpf signal normalized to a lpf signal :/

normalized by*

the lpf would be a constant background sampling

@EwokNightmares you mean optically filtered or electrically filtered?

electrically

just getting signals from a light sensor

if I calibrate from a low background and have knowledge of my intensity from there

i will think about this some more cause my half thoughts are confusing if read :P