A thermal camera is often sold as the answer to a dark site. The pitch is intuitive: ordinary cameras struggle at night, thermal cameras see in total darkness, therefore thermal is the upgrade.
The premise is true. The conclusion does not follow. Thermal cameras do see in total darkness — and they will never tell you who that was.
That is not a limitation of a cheap model, and no amount of resolution fixes it. It is what the technology is.
Two cameras looking at different things
An ordinary camera is a light collector. It gathers visible light that bounced off a scene and arrived at the lens. In darkness there is nothing to collect, which is why a conventional camera needs either ambient light or its own illuminator — the ring of infrared LEDs around the lens that makes night footage look grey and flat. That is what most people mean by "night vision": a camera lighting the scene with light your eye cannot see, then collecting it in the ordinary way.
A thermal camera collects nothing that bounced. It reads radiation the objects themselves emit, as a consequence of their temperature. Axis puts the physical difference plainly: "Conventional cameras work in the range of visible light, which is radiation with short wavelengths. Thermal cameras, on the other hand, detect radiation with mid-length or long wavelengths, which is called infrared radiation."
The operational consequence follows directly: "Since objects themselves emit the heat that thermal cameras detect, thermal cameras are not dependent on visible light and can detect in all light conditions."
So the thermal camera is not seeing better in the dark. It is not using the dark at all. Darkness is simply not an input to it, which is why a thermal image looks identical at noon and at three in the morning.
The sentence that decides whether you should buy one
Here is the manufacturer, describing its own product category: "Unlike conventional cameras, this technology only allows for detection."
And, on people specifically: "While it can determine a person's shape and size, it doesn't allow for clear identification, so privacy is always guaranteed."
Take the first half of that second sentence and set the second half aside for a moment — we will come back to it, because it is a vendor's claim and not a legal finding.
A thermal camera will reliably tell you that a person is there. It will not tell you which person. There is no face in a thermal image, because a face in the thermal band is a warm oval. There is no clothing colour, no logo, no plate, no text. A tattoo does not exist. The distinguishing marks that make an image evidence are all properties of visible light, and visible light is the thing this camera does not collect.
If you are trying to work out whether thermal fits your site, that sentence is the whole test. Ask what you want the footage for.
Where thermal is the right instrument
If the job is knowing that something is in a place where nothing should be, thermal is excellent, and often better than anything else at any price.
A fence line at night. A yard full of equipment. A roof. A rear lot backing onto trees. The failure mode of conventional cameras in those settings is not that they see nothing — it is that they see too much of the wrong thing. Headlights sweep, rain lights up like static in the illuminator, a security light kicks on and blows out half the frame, a plastic bag crosses the lot. Every one of those generates an alert, and a system that cries wolf gets ignored, then unplugged.
Thermal is largely indifferent to all of it. A warm body against a cold yard is an unambiguous signal, and it stays unambiguous through rain, through headlight glare, and through the difference between a floodlit corner and a dark one. Pair it with analytics and you get perimeter detection that holds up across a whole night without a human watching, which is the actual thing most sites need and rarely have.
If the job is identifying who did something, thermal is the wrong instrument and no configuration rescues it.
The design that works: thermal detects, optical identifies
The mature answer on a serious perimeter is not one camera type. It is two, doing different jobs.
Thermal watches the line and decides something is there. That decision is reliable and cheap in false alarms. The alert then drives an optical camera — often one that can pan and zoom — to the spot, where visible light and enough pixels on target produce an image a human or a court can use.
That division of labour is why thermal appears on sites you would not expect to need exotic equipment. It is not there to produce pictures. It is there to be the trigger that is worth believing, so that the camera which does produce pictures is pointed at the right place at the right moment.
If you are quoted thermal as a straight replacement for optical cameras rather than as a layer in front of them, ask what produces the identifying image, and where. If the answer is "the thermal camera", the design has a hole in it.
Two things to check on site before you commit
Point one at a window. Thermal does not pass through ordinary glass the way visible light does. A thermal camera aimed at a window will substantially show you the window, not the room beyond it — which surprises people who assumed they could cover a glazed frontage from inside. It also means a thermal camera cannot be mounted behind glass in a housing, a soffit, or a vehicle windscreen and still work. This is a five-minute test with a demo unit and it settles the question for your specific building better than any specification sheet.
Look at a scene where everything is the same temperature. Thermal images are built from differences. A warm body on a cold night is vivid. The same body on hot asphalt at the end of a summer day, at close to the same temperature, is not. Thermal performance is a function of contrast in temperature, not of darkness, and BC's shoulder seasons produce more of those flat afternoons than a demo in January will show you.
The privacy question, answered carefully
Axis's phrasing — "privacy is always guaranteed" — is a vendor describing a technical property, and it is a fair description of the technology. It is not a legal conclusion, and it should not be repeated to you as one.
The genuinely useful point underneath it is real, though. If a camera cannot produce an identifying image of a person, it collects less about that person than one that can. Where you have a monitoring need in a location that is sensitive precisely because people are identifiable in it, thermal is worth considering on those grounds and not only on darkness grounds. It is one of the few cases where the technically weaker instrument is the more appropriate one.
Whether any given installation satisfies BC's Personal Information Protection Act is a separate question that turns on purpose, notice and proportionality rather than on camera type, and it is the subject of its own article on this site. A thermal camera does not remove a system from PIPA's reach. It changes what that system collects, which is a factor in the analysis and not the end of it.
The short version
Thermal is not a better night camera. It is a detector.
Buy it when you need to know that something is there, across a large or dark or weather-exposed area, with few enough false alarms that people still trust the alerts in month six. Do not buy it expecting to recognize anyone, and do not accept a design where it is the only thing watching a place you would need an identifying image of.
The best question to put to anyone quoting you thermal is the same one that clarifies most security purchases: when this system does its job perfectly, what do I end up holding? For thermal the honest answer is "a reliable alert, at a known time, in a known place." That is worth a great deal. It is just not a picture of anybody.
