Start with what the camera is actually doing
A thermal imaging camera does not see water, pipes or damp. It measures infrared radiation coming off the surface it is pointed at and converts that to an apparent surface temperature, then paints the range across the frame in false colour. Teledyne FLIR, who build most of the cameras in use, put it plainly: thermal cameras cannot see through walls. Point one at a wall and it reads the wall.
The reason it is still one of the best leak tools available is the indirect route. If something behind the surface changes the temperature of the surface by enough, the camera sees the pattern that change makes. A hot pipe leaking under a screed floor warms a plume of screed around it, and the plume reaches the tile. That plume is what you are looking at, not the water.
The two signatures worth knowing
The warm plume
A hot feed or a heating circuit leaking under a floor produces a warm patch that is wider and less regular than the pipe run itself, often teardrop-shaped, with the blunt end at the leak. Run the heating for an hour first, then image the floor cold-to-warm as it comes up. The pipe runs appear as clean parallel lines; the leak appears as the blob that spoils the pattern.
The cold patch
Wet material evaporates, evaporation takes heat, and so a damp area of plaster or plasterboard usually reads slightly colder than the dry wall around it. This is how a cold-water leak shows up at all. It is a much weaker signal than a hot plume, the difference can be a fraction of a degree, and it depends on the room being warmer than the wet material and on air moving over it.
What defeats it
Thermography in buildings is a qualitative method. The British Standard for detecting thermal irregularities in building envelopes, BS EN 13187, describes it that way and sets out the conditions a survey needs, including a genuine temperature difference across the fabric. That requirement is the whole story of when leak thermography fails.
- A cold leak in a cold wall shows nothing. A mains cold pipe weeping into an unheated hallway wall in a cold house gives the camera no differential to find. The classic false negative is a survey run in an empty, unheated property, which is exactly when people tend to discover a leak.
- Depth and insulation flatten the signal. Thick screed, an insulated floor build-up or a chase packed with mineral wool spread the heat out before it reaches the surface. A pipe 25mm below tile shows sharply; the same pipe under 100mm of insulated floating floor may show nothing.
- Underfloor heating hides its own leaks. On a floor where the whole slab is warm, a warm plume has nothing to stand out against. Underfloor heating leaks are often better found cold, by draining the circuit and using tracer gas, or by imaging while the circuit is filled with cold water against a warm slab.
- Emissivity and reflections lie. Different materials radiate differently: brick and plaster are efficient, polished metal and glass are not. A shiny tile, a foil-backed board, a radiator across the room reflecting in a gloss floor, all produce hot and cold marks that have nothing to do with water. Interpreting these out is the skill, and it is why a thermal image on its own is not evidence.
- It cannot tell you what the water is. A warm damp patch on a bathroom wall could be a leaking feed, a failed shower seal, or condensation running behind a splashback. The camera shows a pattern; the diagnosis comes from combining it with a moisture map and a pressure test.