We earn a commission if you buy through links on this page, at no extra cost to you.

The DORI Pixel-Density Decoder

DORI — Detect, Observe, Recognise, Identify — is a real standard, IEC/EN 62676-4, that defines exactly how many pixels per meter (px/m) of a subject a camera needs to deliver each task. Identification needs 250 px/m — roughly 10x the 25 px/m needed just to Detect that someone is there. This page decodes the thresholds and shows the calculation to estimate your own setup.

The one thing to know: DORI stands for Detect, Observe, Recognise, Identify — four escalating tasks defined by the international video-surveillance standard IEC/EN 62676-4, each requiring a specific minimum pixel density on the target (measured in pixels per meter, px/m): Detection needs about 25 px/m (something is there), Observation about 62.5 px/m (general characteristics), Recognition about 125 px/m (probably a known person), and Identification about 250 px/m (identity beyond reasonable doubt). Most consumer camera marketing never mentions which DORI level its advertised range refers to.

License: CC BY 4.0 — free to use with attribution

This dataset is released under a Creative Commons Attribution 4.0 International license. You may use, share, and adapt it for any purpose, including commercially, as long as you provide attribution.

How to cite

Night Vision Decoded. "DORI Pixel-Density Decoder" (Version 1.0), July 30, 2026. https://nightvision-decoded.pages.dev/dori-pixel-density-decoder/

The DORI thresholds

DORI levelMin. pixel densityPractical meaning
Detect (D)25 px/mConfirm a person or vehicle is present and distinguish it from the background.
Observe (O)62.5 px/mDiscern general characteristics — rough shape, posture, clothing color.
Recognise (R)125 px/mDecide with reasonable confidence whether an already-known person is the one shown.
Identify (I)250 px/mEstablish identity of an unknown subject to a level suitable as evidence.

Source: IEC/EN 62676-4, the international video-surveillance-system standard defining performance requirements by pixel density on target. Published in Europe as EN 62676-4 and in Australia/New Zealand as AS/NZS 62676.4.

How to estimate your own camera's DORI level (derived calculation)

This is a derived estimate from the published standard, not a lab measurement of any specific camera. The formula: divide your camera's horizontal resolution in pixels by the real-world width, in meters, that the frame covers at your subject's distance. Compare the result to the table above.

Worked example: a 2560-pixel-wide (roughly 4MP) camera whose field of view spans 12 meters at your driveway's distance delivers about 2560 ÷ 12 = 213.3 px/m at that distance — above the 125 px/m Recognition threshold but below the 250 px/m Identification threshold. To reach Identification at that same distance, you would need roughly double the horizontal resolution, a longer lens (narrower field of view covering less real-world width per pixel), or to be roughly half the distance.

Field of view width at a given distance depends on the lens's focal length/angle of view, which is a separate spec from resolution — a camera with a very wide-angle lens can have high resolution but still deliver low pixel density on a subject because that resolution is spread across a much wider real-world scene.

Why this matters more than "advertised IR range"

A camera can genuinely Detect motion at 100 feet in the dark (only 25 px/m needed) while being completely unable to resolve an Identification-quality face at that same distance (250 px/m needed) — a roughly 10x gap in required resolution that most marketing never surfaces. See Can My Camera Identify a Face at Night? and What Does IR Range Actually Mean?

Caveats

Methodology & versioning

Version 1.0 — published July 30, 2026. Pixel-density thresholds are drawn from the published IEC/EN 62676-4 standard and cross-referenced against professional CCTV-industry technical explainers (Axis Communications' published pixel-density white paper, and CCTV-calculator vendor documentation) that cite the same standard. We did not conduct our own lab testing. Corrections: if we find an error in the threshold figures or the standard is revised, this page and CSV are updated in place and the version/date bumped — see How We Evaluate for our full editorial policy.

Machine-readable download: the full table is available as CSV at /dori-pixel-density-decoder.csv (version-pinned copy at /dori-pixel-density-decoder-v1.0.csv).

Frequently Asked Questions

Detect, Observe, Recognise, Identify — four escalating surveillance tasks defined by the international video-surveillance standard IEC/EN 62676-4, each requiring progressively more resolved detail (pixel density) on the subject.

Per IEC/EN 62676-4: Detection needs roughly 25 pixels per meter (px/m) of the subject — confirming something is present. Observation needs roughly 62.5 px/m — general characteristics like posture or clothing color. Recognition needs roughly 125 px/m — deciding with reasonable confidence whether a known person is the one shown. Identification needs roughly 250 px/m — establishing identity with a level of detail suitable as evidence.

It's a real standard — IEC 62676-4, published in Europe as EN 62676-4 and in Australia/New Zealand as AS/NZS 62676.4. It predates most consumer smart-camera marketing and was developed for professional CCTV system design, which is why the pixel-density math is well-defined even though almost no consumer camera box mentions it.

Estimate the pixel density your camera actually delivers on a subject at your intended distance (horizontal resolution in pixels divided by the real-world scene width in meters at that distance), then compare it to the DORI thresholds. See our DORI Pixel-Density Decoder for the full worked calculation — treat it as a derived estimate, not a certified test of your exact hardware.