Sounds like a lot of trial and error ahead! As long as particulate solids are fine enough and dispersed enough, your imagers spot size will probably cause the imager to ignore this particulate to a large degree. Moisture in the air will possibly prove to be the second biggest annoyance. According to the ASNT Non Destructive Handbook, Third Edition, Volume Three, Chapter Three, you will find that water is a big attenuator in most spectrums used in short and long wave IR. The largest gap in (least)effect of moisture in the air on transmittance is in the 5.5 to 7.25 micron range, indicating least impact from the moisture using a long wave imager. You will still have to deal with scatter caused by large moisture particulate. The largest limiting factor might simply be the imagers spot size. Be careful of the IFOV method for determining your imagers spot size, it can be as much as 5 times larger than actuality(typically twice).
I have had to do something similar in that I was requested to help in tracking/determining heat at a engine drive shaft boring bit on a shaft fabrication line. Excess heat at the bit tips was being blamed for early dulling and shaft overheating and temper faults at boring points.
Good Luck, I would appreciate any feedback!