Thermal camera spot size and IFOV

Every pixel of a thermal camera looks at a cone of the scene. The angular width of that cone is the instantaneous field of view, and it follows from the lens and the detector: IFOV = 2·atan(tan(HFOV/2)/N) for a horizontal field of view HFOV spread across N pixels. Projected onto a flat target at distance d, one pixel covers a spot 2·d·tan(HFOV/2)/N across. A 24° lens on a 320-pixel detector gives 1.33 mrad per pixel, which is 6.6 mm at 5 metres.

The trap is that one pixel on the target is not enough to measure its temperature. Optical blur, detector crosstalk and the finite slit response of the system mean a single pixel sitting on a small hot object still integrates a large share of its surroundings, and the camera reads low. Manufacturers therefore publish a measurement spot size, commonly three times the geometric IFOV, and thermographic procedures require the target to be at least that big. Take a thermographic temperature only when the target comfortably exceeds the measurement spot; anything smaller is a detection, useful for saying that something is hot, not for saying how hot.

Turned round, the same geometry gives the maximum standoff for a target of known size — the number that decides whether a bolted connection can be surveyed from the walkway or needs a lift. Halving the distance halves the spot, as does doubling the pixel count or fitting a narrower telephoto lens; note that a telephoto buys resolution at the cost of coverage, so the survey takes proportionally longer.

These figures are geometric best cases. Focus is critical — a defocused image inflates the effective spot far beyond the geometric value — and off-axis pixels see slightly larger spots. Intervening glass, mesh guards, plastic covers and heavy smoke break the measurement entirely rather than merely enlarging the spot, because they are not transparent in the 8–14 µm band.