Seeking advice on the most accurate low end products

[Sad Bison]

Hi all,

I'm currently looking into using thermal imaging in some biology research and am trying to work out the best low end/budget product that will give the highest accuracy in temperature readings. We plan to use the thermal imager to measure the temperature of various microhabitat surfaces as well as body surface temperatures of reptiles. I would expect the temperatures we're measuring would mostly fall somewhere between 5 - 60 degrees C.
We had initially looked at the FLIR One cameras as the hybrid thermal/visual output looks like it would make it quite easy to see exactly what features we're looking at (and measuring) without needing to spend extra for a higher resolution sensor. However, with a reported error of +/- 3 degrees C this would probably not be accurate enough.
Unfortunately, I don't yet know how much funding we'll be able to secure but at the moment I can't see anything over $1000 being a realistic option.

So I have two main questions I'm hoping someone might be able to help me with:
a) Are there any similar(ish) low-end products which actually have a lower temperature measurement error than the FLIR One?
b) Has anyone here measured the accuracy of any of these products themselves, and have any tips or tricks on how to reduce measurement error and get the most out of these cheaper imagers?

Cheers,
SB

[Chanc3]

This article from FLIR is definitely worth a read: http://www.flir.com/science/blog/details/?ID=74935 Basically, the accuracy value is taking into account many many factors that can affect the reading - mainly the operator.

We've got 2x blackbody calibration sources here and haven't had anything over 1°c out at the lower temperatures - can't really get much better than that.

[Bill W]

I would suggest that your biggest problem will be compensating for emissivity.  It is one thing getting 1°C accurate / repeatable with a black body, but with real world emissivity / reflection to account for it is going to get tricky.  The next biggest error is likely to be changing camera / lens temperature.

I'd even suggest a lizard might change its' emissivity.

Could you get a known (or measurable) temperature block or plate in the scene ?  That would take out most of the camera 'offset' errors.

An alternative to a camera is the cavity probe, see 'QHi Exertherm' as these are proof against emissivity.

Bill

[Sad Bison]

Thanks for the responses! Definitely all helpful points.

The article you linked to Chanc3 was an interesting read - I'm not sure how I didn't stumble across it before! Seems like the cheaper options may still be usable then.

Emissivity could be a problem - I will definitely need to look into estimating emissivity of the various materials in the environment we'll be measuring - but indeed the lizards do change their reflectivity at different temperatures, so from what I understand that would mean emissivity too. But good idea! - getting an object of known temperature in the scene should be doable - or independently measuring something in the scene as a reference.