0
$\begingroup$

Not sure if this is the right place to be asking this, but it's my best bet.

So, I have a video captured from an infrared camera similar to the youtube video here. Can I get temperature information by analyzing the pixel data of any given frame? Also, if this needs prior calibration, how would I go about doing that?

|improve this question
$\endgroup$

migrated from stackoverflow.com Dec 26 '17 at 20:29

This question came from our site for professional and enthusiast programmers.

  • 1
    $\begingroup$ You better first ask about the physical machanism of capturing temperature data at the IR range by an IR CMOS sensor (at electronics.se etc), then come back here with the formula, to ask how to implement it... $\endgroup$ – Fat32 Dec 26 '17 at 20:52
  • 1
    $\begingroup$ the youtube video you show is barely IR, essentially a conventional visible sensor pushed to the limit of red. An actual IR sensor has different sensitivity. An automobile shows pronounced heat associated with the exhaust system and windows are opaque. Glass used for conventional cameras are useless in IR. Galium Arsenide is typically used in IR lenses. $\endgroup$ – Stanley Pawlukiewicz Dec 26 '17 at 21:28
1
$\begingroup$

There is no direct relation between the emitted intensity and temperature, as this depends on the body color and many other factors.

Anyway, you can get an approximation by calibration. For this you use a reference object and heat it to known temperatures. You will draw a calibration curve (pixel intensities as a fuction of temperature) and possibly map a simple model (polynomial) onto it.

|improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.