[exiftool] Question about parameters R1, R2, B, F, O / Tag values in FLIR images

[vitorpaes]

Hello there,

I am working with some SC660 Flir images on exiftool and i have some questions about the information extracted on the program:

1 - What is the physical definition of the parameters R1, R2, F and O. Are they just an aproximation of the Plancks Law for blackbody radiation?
2 - exiftool(-k) returns some raw values tagged as Raw Value Range Max and Raw Value Range Min which leades to the interpretation that they are the values for the range of operation. However, when I convert them to temperature, it returns 150 and -60 but the range of operation is 120 -40.  What does that mean ? Could it be that the range of operation is indeed -40 120 but the saturation points are -60 150?

Thank you.

[JimM]

1) In Exiftool, those parameters (R1, R2, B, F, and O) are prefixed by the name Planck. I think Phil Harvey (of Exiftool fame) did that as a convenience for grouping tags. However, they aren't directly related to Planck's Law. Rather, they are calibration constants and global gain and offset parameters for FLIR cameras.
Looking at this further, some of the parameters like B and R are related to Planck's Law. Other parameters like R2 and O are camera gain and offset. The posts by tomas123 on the Exiftool forum have the details.

2) Don't know the answer to that.

If anyone else has further information on this, please post.

[JimM]

Looking further into this (if anyone is interested), I found a paper by Glenn Tattersall, attached below, that references these parameters. At the bottom of page 14 of the paper, there is an equation similar to Planck's law: s = R/(e^B/T - F) where s is the radiation signal being detected by the camera. If you solve for T (temperature in Kelvin), at top of page 15 you get: T = B/ln((R/s) +F) where ln is natural logarithm.

FLIR has modified this by breaking R into two parts: R1 and R2, where R = R1/R2. Furthermore, there's an offset term O, so you have s + O. So then the FLIR equation is:
T = B/ln((R1/R2)(1/(s+O)) + F). This is the basic equation FLIR uses for calculating temperature if the effects of emissivity, reflected temperature, or atmospheric losses are not included.

For further information, there's also a book: Minkina and Dudzik: "Infrared Thermography: Errors and Uncertainties",  ISBN-13: 978-0470747186 It costs $107

Edit to add: You can download a PDF of the Minkina, Dudzik book for free. So, you don't need to spend $107 unless you really need a hard copy.

http://dr.library.brocku.ca/bitstream/handle/10464/8607/Tattersall%20Thermography%20CBP-A%20Accepted.pdf?sequence=1&isAllowed=n

[eri0o]

Hey JimM, this is a message from tomas123 with a lot of information that may be useful for you.

https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg342072/#msg342072

[JimM]

Hey JimM, this is a message from tomas123 with a lot of information that may be useful for you.

Yeah, I'm aware of the tomas123 posts on this subject. I was just trying to answer the question from the OP vitorpaes, about where the R1, R2, B, F, O parameters came from. In the Minkina and Dudzik references, they actually only mention R, B, and F. But they also mention Global Gain and Global Offset. In the tomas123 and Phil Harvey (ExifTool) rendition, I think R2 = Global Gain and O = Global Offset. However, does that mean R = R1 or R = R1/R2 ? I haven't been able to resolve this.

In other posts from tomas123, he references a total of 15 parameters in the FLIR metadata that you use in order to calculate temperature when accounting for emissivity, reflected temperature, and atmospheric losses. He also posted a spreadsheet where he has all of the calculations, and he shows for a 16-pixel test image that he gets the same temperature values (to about 5 decimal places) as FLIR tools for object distances of 0 meters and 100 meters. So, although I can't completely resolve the equations above, the proof is in the pudding on the tomas123 temperature calculations. In looking at scraps of information from FLIR documents, there is not consistency in how to calculate temperatures. It might be that things changed over the years, or there might be alternative methods for calculating temperatures, or different methods for different cameras.