Exploring Photo-Sensor Measurements for Brightness Change Evaluation

[Kamuro]

Hello,

I'm using photo-sensors from commercial and industrial cameras to document an experiment. Afterwarde I evaluate the change in brightness during the test of a fixed segment of every image I took. The ambient light condition is stable and the images are in raw files, but I'm unsure how raw they are.

I read that in photometry, the radiant power of each wavelength is weighted by a luminosity function, because they want to mimic the sensitivity of the human eye to different wavelengths.

Does this happen here, too? I mean, I know that the sensors are reading absolute values, but are the absolute values weighted by a luminosity function in the camera afterwards? Or do the images contain the absolute values and our eye does the weightening?

In the end, I just want to know if the pixel of the images I reading out are in absolute or weighted values. I could not find anything so far on the web. I'm grateful for every help I'm getting!

regards
Jim

 

[Andy Resnick]

In the end, I just want to know if the pixel of the images I reading out are in absolute or weighted values. I could not find anything so far on the web. I'm grateful for every help I'm getting!

I think you should ask the sensor manufacturer for a 'quantum efficiency' curve, that will tell you the spectral weighting function of the sensor. I doubt the QE curve will match a photometric curve...

 

[russ_watters]

Interesting question. I would think that for most higher end commercial/industrial/scientific cameras you are getting the "raw" output because it is more useful for post-processing to have all the data.

Also, it would be logically flawed to filter the light according to human vision when recording the image because then if you display the image it looks double-filtered -- unless you are specifically trying to measure something according to human perception (which is itself variable and imprecise).

That said, it is important to remember that pretty much all cameras are inherently monochrome and color is typically added in post-processing based on selective filtering during image acquisition.

 

[sophiecentaur]

@Kamuro
If you are discussing Photometry then the following Wiki statement is relevant:
"Photometry is the science of the measurement of light, in terms of its perceived brightness to the human eye."
So that answers one of your questions.

Or do the images contain the absolute values and our eye does the weighting?

The eye cannot be relied on to do the weighting.
The logic is that the eye / brain could only do the weighting precisely if the precise spectrum of the original scene were presented to it. This never happens in any imaging system. A monochrome camera / photometer will have a weighting filter and it would, presumably, have a monochrome display of some sort of colour. If the display is 'white' than the particular white point would need to be specified. Clearly not practical. A colour camera uses three colour analysis with three filters and the display uses three primaries for the display. The analysis and display are aimed at giving as good subjective colour match as possible (it cannot reproduce the spectrum of the scene) - the eye can only work with that.
A photometer will give you a number (perhaps three numbers) - not a picture of a colour.

In addition, the eye's sensitivity / perception is not linear and it is time dependent (dark adaptation etc.). If a photometer is to give the perceived brightness for an observer, there would be a lot of calibration needed, to take actual light levels and viewing conditions.