Cameras, Racks and Power · #77

Online Camera Pixel Density Calculator

Shows with pixel density whether a camera will detect or recognise a person at a given distance.

  • Data is processed in your browser and not sent to a server

Tool

E.g. 1920, 2560, 3840.

HFOV from datasheet (degrees).

Known width; instead of angle.

What does this tool do?

'8 MP camera' says little: with a wide lens there may not be enough pixels to recognise a face at 20 m. The deciding value is horizontal pixels per metre in the scene. The tool compares PPM with the IEC 62676-4 DORI thresholds (25/62/125/250 px/m) and, with a field of view, gives the furthest distance for each level.

How to use it

  1. Enter the camera's horizontal pixels (e.g. 2560 for 4 MP).
  2. Enter either horizontal field of view and target distance, or the scene width at the target.
  3. Press Run.

Fields

Horizontal resolution
E.g. 1920, 2560, 3840.
Horizontal field of view
HFOV from datasheet (degrees).
Target distance
Scene width
Known width; instead of angle.

Worked example

InputValue
Horizontal resolution2560
Horizontal field of view90
Target distance10

Expected result: 2560 px, 90° FOV, 10 m: scene 20 m wide, 128 px/m → recognise. Identify up to 5.1 m, observe up to 20.6 m.

If face identification is required at the entrance, place the camera within 5 m or choose a narrower (varifocal) lens. Use the Lens Focal Length Calculator to pick the lens.

Second scenario

InputValue
Horizontal resolution3840
Scene width6

Expected result: 3840 px over a 6 m wide door: 640 px/m, more than twice the identify threshold.

Method and assumptions

Scene width = 2 × distance × tan(angle / 2). PPM = horizontal pixels / scene width. PPF = PPM × 0.3048.

DORI (IEC 62676-4): detect 25, observe 62, recognise 125, identify 250 px/m. Level distance = (pixels / threshold) / 2 / tan(angle / 2).

Using the result and troubleshooting

Target identify or recognise for critical points (entrance, till, plates); observe is enough for general areas. Check that IR illumination range covers this distance at night.

Limits and common mistakes

  • Lens distortion, compression, focus, light and night performance are not considered; real image quality may be below the PPM.

Frequently asked questions

How many pixels are needed to read plates?

Vendors usually recommend at least 130–150 pixels across the plate for LPR (≈ 250–300 px/m); angle and speed also matter.

Related tools and service

Related service: Camera Positioning and Blind Spot Analysis

CCTV online tools · See all online tools