Identify anyone. It shows a warm shape, not a face, clothing colour, plate or logo - and no processing recovers that detail.

Thermal images are formed from emitted heat, and heat carries none of the fine detail the eye uses to recognise a person.

That makes thermal a detection sensor rather than an evidential one.

Evidence requires a visible-light camera, and the two are almost always specified together.

It also cannot see through glass, because glass is opaque to the wavelengths involved.

Setting the expectation early avoids the common disappointment of a site that bought thermal cameras and cannot identify anybody on the recordings.

Considerably further than a visible camera at night - but the detection range is far longer than the range at which anyone can tell what they are looking at.

Manufacturers quote separate figures for detection, recognition and identification, and the difference between them is large.

Detection means something warm is there. Recognition means it is probably a person rather than an animal. Identification is not achievable with thermal at all.

Range depends on the detector resolution and the lens, and a longer lens narrows the field of view proportionally.

Atmospheric conditions reduce range: heavy rain, dense fog and high humidity all attenuate thermal radiation.

Specify against the detection distance actually required, since longer-range units cost substantially more.

Because the image depends on heat rather than light, so the things that fool an analytic on a visible camera simply are not present.

Shadows, headlights, sunrise, floodlights switching on and reflections off wet ground cause most outdoor analytic false alarms.

None of them appear in a thermal image, because none of them are heat sources.

A person is a consistently strong signal against a cooler background, day and night.

This is the main reason thermal is used on perimeters even where some visible-light capability exists.

It is not perfect: solar-heated surfaces, vehicle exhausts and animals all produce thermal signatures that need filtering.

Through smoke and light fog, yes, and through thin foliage to a useful degree - but not through glass, and dense fog attenuates it.

Smoke particles are largely transparent at thermal wavelengths, which is why fire services use thermal imaging.

Light fog and haze are penetrated far better than by visible light, though dense fog with large water droplets does attenuate the signal.

Foliage is partially penetrated: a warm body behind a hedge often shows where a visible camera sees only leaves.

Glass is opaque to thermal, so a camera cannot look through a window, and a person behind glazing is invisible.

Rain attenuates and also cools surfaces, reducing the contrast the image depends on.

Yes - in several jurisdictions thermal imaging equipment above certain performance thresholds is export-controlled, which affects international procurement.

Controls typically key on detector resolution and frame rate, so higher-performance cameras are more likely to be restricted.

This affects shipping across borders, re-export, and sometimes the end user and the stated application.

Suppliers generally know the position for their own products, and it should be confirmed in writing before an order is placed.

For multinational projects, expect different model availability in different territories rather than one global part number.

Spares and warranty replacements are subject to the same rules, which is worth establishing at the outset rather than during a failure.