Home / Guides / CCTV Lens Selection Guide

Field guide · CCTV & security

Choosing the Right CCTV Lens: Coverage vs. Identification Distance

The most common lens mistake isn't a wrong formula — it's picking a wide lens because it "covers more," then finding out during an actual incident that nobody in the footage can be identified.

Coverage and detail are opposite trade-offs

A wider lens (shorter focal length) spreads the same sensor resolution across a bigger scene — more area covered, fewer pixels per meter of that area. A narrower lens (longer focal length) does the opposite: less area, but every object in frame gets more pixels. Neither is "better" — they're suited to different jobs, and the mistake is picking one without checking what the resulting pixel density actually gives you at the distance that matters.

Formula: Horizontal field of view = 2 × atan(sensor width ÷ (2 × focal length)). Pixel density at a distance = horizontal resolution ÷ scene width at that distance.

Run those two together and you get the number that actually matters on site: how many pixels per meter fall on a person standing at your entrance, your gate, or the far end of your parking lot. The Camera Field of View Calculator does this automatically from sensor size, focal length, resolution, and distance.

DORI levels: the standard's own yardstick

IEC 62676-4 defines pixel-density levels known as DORI (Detection, Observation, Recognition, Identification), and manufacturers such as Axis publish the same four thresholds. They come from a published standard, not marketing:

ClassificationPixel densityWhat you can actually do with it
Identification≥ 250 px/mConfirm exactly who a person is, or read a plate
Recognition≥ 125 px/mRecognize someone you already know, or a vehicle type/color
Observation≥ 63 px/mFollow what a person is doing and notice general details, but not recognize or identify them
Detection≥ 25 px/mNotice that a person or vehicle is present at all

Below 25 px/m, the footage is only good for general scene awareness — motion happened, but you can't say who or what caused it. This is the gap that catches technicians most often: a camera spec sheet's megapixel count says nothing about which of these four tiers you're actually getting at your specific install distance.

The mistake: sizing the lens for the room, not the job

A common install pattern: pick a 2.8mm lens because it "gets the whole room in frame," mount it, and move on. That lens might genuinely cover the whole space — and still leave every face in it below detection grade the moment someone is more than a few meters from the camera. The lens did its job of showing the area; it never had enough resolution per meter to show who was in it.

The fix isn't a longer lens everywhere — a narrow lens at a doorway loses the wide coverage you need to see someone approaching. It's matching the lens to the job at each specific mounting point: wide coverage where general awareness is the goal (a parking lot overview), longer focal length where identification is the goal (an entrance, a till, a gate).

Rule for the field: before finalizing a lens, calculate the pixel density at the farthest point where identification actually needs to happen — not the farthest point the lens can physically see. Those are rarely the same distance.

Sensor size matters as much as focal length

Two cameras with the same focal length but different sensor sizes have different fields of view — a larger sensor behind the same lens sees a wider angle, not a narrower one. When comparing lens specs across camera models, focal length alone is not comparable unless the sensor size is the same; this is why the field of view calculator asks for both.

Making the call

Identify the distance where each camera actually needs to deliver identification, recognition, or just detection. Run sensor size, focal length, resolution, and that distance through the Camera Field of View Calculator to see which tier you land in — and, if it's short, whether a longer lens or a higher-resolution sensor closes the gap more cheaply. Treat the result as a best case: lens distortion, compression, and low light all eat into it in the field.

Sources

The Detection, Observation, Recognition and Identification levels (25, 63, 125 and 250 px/m) are as published by Axis in the papers above. IEC 62676-4 itself is a paid standard and its text was not used here. Axis notes that an IEC 62676-4:2025 edition exists; the levels on this page are the 2014-based ones from the Axis papers.

Found something wrong on this page? Send feedback →