infra red temperature measerment

[Mikosim]

Greetings to all.

Сan anyone have experience in this matter? Necessary to conduct video surveillance of the object and measure the temperature of each pixel or group of pixels in the frame. Which camera is desirable to use? What additional software to use?

[Klemen]

Hello Mikosim,

for measuring themperatures the spectral band ranging from ~ 0.7 um to 14 um is used. Lower the wavelenght, the higher temperatures can be measured. I would suggest a typical system that ranges from 8um - 14um. That enables you to measure the temperatures from approximately -50° to 1000°. This spectral band is used to measure non-transparent bodies with high emissivity (human skin is 0.98).

About the camera - you need to think about the sensitiviy or NETD (basically signal to noise ratio). Here is a quick explanation:

http://irinformir.blogspot.com/2012/05/what-does-sensitivity-netd-mean-when.html

Cameras usually have automatic (or manual) calibration that reduces this noise as it grows during the measurements.

You should also consider the distance from the measured object and thus taking the appropriate lens type. Usually, when you look at the IR camera specifications, all the data needed to calculate the imging parameters are present there (you can then calculate on how many pixels, the object will be imaged).

Next, you should consider if you want to buy uncalibrated or calibrated camera. Most cameras (as far as i know) are already calibrated. You should take care if you plan to change the optics of the camera, new calibration should then be performed, because the pixel's response will not be the same. I suggest buying the right type of camera in the start (calibration bodies cost some thousands of euros).

About the software you should only take care that it is compatible with the directshow protocol. You can then build your own specific software. Usually cameras also have an additional port (rs232) for two-way communication (settings, etc.).

Also, pay some attention to the camera drivers (x64 or x86).

Hope this helps a bit.

Best regards,

K

[df86]

We normally use FLIR Cameras for that type of projects.

The cameras have different types of communication protocols but the most common one is GigE using GeniCam protocol. This way you can have the Images transferred to labview and manipulate it using Ni Vision development toolkits. There are also drivers and softwares available in order to communicate with these cameras if you do not want to go through the hassel of creating your own custom software.

Sorry about the self promotion but you can visit our website, http://www.movitherm.com/. there you can see the range of the cameras available. You don't only have to deal with technical issues since these Cameras are ITAR controlled, so speaking to someone regarding your application and location might not be a bad idea.

[Mikosim]

Thank you so much for your help gentlemen! I will clarify the problem: the spectral diapasone in which I want to measure: 2 - 5 micron and distance to the subject 500 - 4000 meters. it is much difficult situation?

[Klemen]

Hello,

the distance seems too much for me. What about rain, fog and other atmospheric conditions? You would have to consider all of these factors to get the correct reading (you need to measure the temperature, right?). About the optical part of the system, you should have no problems, since there are lenses capable of providing reasonably good image at such distances (long-focus lenses). I would say you could get a good accuracy at the distance up to 100 m. This is just my opinion. I think you should contact a company that deals with IR measurements and get the information from experts that deal with this daily.

What are you measuring. 2-5 um are commonly used for measuring the temperature of gasses (CO2, CO), metals with low emmisivity or maybe glass surfaces.

Best regards,

K