A potential source of confusion… 

Just as you can see your refection in a window, you can see it in a thermal image too. Thermal reflections in glass or metal surfaces  cause misleading temperature readings (as can their low emissivity).

If you’re not sure whether you’re seeing thermal reflections, try moving left and right, while watching the camera screen: reflections move differently from features seen through the glass.

At infrared wavelengths, glass is often more mirror-like than for visible light, depending if the glass has special coatings.  Low-e glass is coated to be more reflective (low emissivity) at these wavelengths. This makes these windows more thermally efficient, because radiant heat from warm things in a room (like you) is reflected back, rather than transmitted to the outside.

Be cautious interpreting thermal images of glass:

This image was taken on a very frosty morning, just after dawn. The windows are single-glazed with added secondary glazing and both rooms were unheated. From the spot temperatures:

Sp1 -8°C:  the solid brick wall between the 2 windows.
SP2 -12°C:  the upper window actually looses more heat through it than through the wall. It appears so cold because of the reflected clear sky.
SP4 -60°C:  the clear sky is probably even colder, but this is the lowest the Flir One can measure.
SP5 -4.1°C:  the lower window has reflections of the warmer(?) house across the street.

So beware of reflections in glass or metal surfaces! Look to see what is visibly reflected and consider its likely thermal effect.  In general don’t trust glass temperatures.

The glazed front door of the house on the left appears to be hot (Sp1= 6.7 °C) and so is apparently leaking heat. Consider carefully if what you see in the glass could be reflected heat instead –  maybe reflections of hot houses on the camera-side of the street? In fact, when checked, those houses were less hot.

Below the letterbox, the wooden door isn’t glazed and so its temperature (Sp2= 5.5°C) is more reliable. This suggests the glazed door panel above is loosing a bit more heat than the lower wooden panel, which seems consistent.

So this single-glazed door does seem to be very poorly insulated, losing significant heat. This could be confirmed with PVC tape on the glass (see below).

In Practice
Beware of reflections in thermal images of any glazing (or bare metal).

If you need to know glass temperature, stick on some PVC electrical insulating tape and measure its temperature by thermal imaging after it’s equilibrated  (maybe 30 seconds).

<- TI Home                                                                                                        Emissivity ->