Quick Answer

The resolution you need depends on how wide the scene is and how much detail you need in it, not on the megapixel label. IEC 62676-4:2014, as summarized by Axis, sets 125 pixels per meter for recognition and 250 pixels per meter for identification. A scene 20 feet wide needs 762 horizontal pixels for recognition and 1,524 for identification, so a 2MP (1920 px) camera covers both. A scene 40 feet wide needs 1,524 and 3,048 pixels, which points to 2MP / 1080p for recognition and 8MP / 4K UHD for identification. These are center-of-image figures for a simplified model, so add a margin.

This page answers the buying question, “which resolution do I need for this scene?” For the pixel dimensions of each megapixel label, see the security camera resolution comparison. For a plain-language explanation of 1080p, 2K, and 4K, see security camera resolution explained. For distance from a specific lens, use the lens distance calculator.

The Pixel Density Rule Behind the Numbers

Axis explains that the pixel density model starts from the number of pixels needed across a human face, then converts that into pixels per meter or per foot by assuming an average face is 16 cm (6 5/16 in) wide. The four levels come from IEC 62676-4:2014. Axis states that they apply when human operators interpret visual video.

Level (IEC 62676-4:2014, via Axis)What Axis says it meansPixels across a facePer meterPer foot
DetectionIt is possible to determine whether or not any individual is present.4 px/face25 px/m8 px/ft
ObservationIt is possible to determine how many people are present and to see characteristic details of individuals, such as distinctive clothing.10 px/face63 px/m19 px/ft (rounded: 20)
RecognitionIt is possible for a viewer to determine whether or not an individual shown is the same as someone they have seen before.20 px/face125 px/m38 px/ft (rounded: 40)
IdentificationIt is possible to identify an individual.40 px/face250 px/m76 px/ft (rounded: 80)

IEC 62676-4:2014 lists px/m. Axis datasheets use the exact conversions (19, 38, and 76 px/ft); the rounded 20, 40, and 80 px/ft values are common in practice. Axis notes that IEC 62676-4:2014 usually recommends at least 40 pixels across a face to identify someone, and says a higher density can add a safety margin in poor light or when people are not facing the camera. The tables below use the px/m values.

The rule gives you one formula: horizontal pixels needed = scene width in meters x pixel density (px/m). To work in feet, multiply feet by 0.3048 first.

Resolution Needed by Scene Width

Each cell is the scene width multiplied by the density, followed by the smallest resolution from our list (2MP / 1080p at 1,920 px, 4MP at 2,560, 5MP at 2,592, 8MP / 4K at 3,840, and 12MP at 4,000 pixels wide) that reaches it. These are calculated values, not measurements.

Scene widthObservation (63 px/m)Recognition (125 px/m)Identification (250 px/m)
6 ft (1.8 m)115 px
2MP / 1080p
229 px
2MP / 1080p
457 px
2MP / 1080p
10 ft (3.0 m)192 px
2MP / 1080p
381 px
2MP / 1080p
762 px
2MP / 1080p
15 ft (4.6 m)288 px
2MP / 1080p
572 px
2MP / 1080p
1,143 px
2MP / 1080p
20 ft (6.1 m)384 px
2MP / 1080p
762 px
2MP / 1080p
1,524 px
2MP / 1080p
30 ft (9.1 m)576 px
2MP / 1080p
1,143 px
2MP / 1080p
2,286 px
4MP
40 ft (12.2 m)768 px
2MP / 1080p
1,524 px
2MP / 1080p
3,048 px
8MP / 4K UHD
60 ft (18.3 m)1,152 px
2MP / 1080p
2,286 px
4MP
4,572 px
More than 12MP
80 ft (24.4 m)1,536 px
2MP / 1080p
3,048 px
8MP / 4K UHD
6,096 px
More than 12MP
100 ft (30.5 m)1,920 px
2MP / 1080p
3,810 px
8MP / 4K UHD
7,620 px
More than 12MP

“Scene width” means the width of the area the camera sees at the distance where the person stands, not the width of the whole image at the far end. A scene that needs more than 4,000 pixels is beyond a 12MP camera at that level of detail, so use more than one camera or move the camera closer.

Megapixel Calculator

Enter the width of the scene at the spot you care about, choose a level, and read the horizontal pixels required. The calculator uses the formula above.

Enter a scene width to see the horizontal pixels required.

Widest Scene Each Resolution Can Cover

This is the same arithmetic in reverse: the camera’s width in pixels divided by the density. For example, a 1,920 px wide image divided by 250 px/m is 7.7 m, or about 25 ft.

ResolutionDetection (25 px/m)Observation (63 px/m)Recognition (125 px/m)Identification (250 px/m)
2MP / 1080p
1,920 px wide
252 ft (76.8 m)100 ft (30.5 m)50 ft (15.4 m)25 ft (7.7 m)
4MP
2,560 px wide
336 ft (102.4 m)133 ft (40.6 m)67 ft (20.5 m)34 ft (10.2 m)
5MP
2,592 px wide
340 ft (103.7 m)135 ft (41.1 m)68 ft (20.7 m)34 ft (10.4 m)
8MP / 4K UHD
3,840 px wide
504 ft (153.6 m)200 ft (61.0 m)101 ft (30.7 m)50 ft (15.4 m)
12MP
4,000 px wide
525 ft (160.0 m)208 ft (63.5 m)105 ft (32.0 m)52 ft (16.0 m)

Notice that the width a camera can cover grows with horizontal pixels, not with the megapixel number. 2MP to 8MP is four times the megapixels but only twice the width. The 4MP and 5MP sizes in this table differ by only 1.25% in width.

Worked Examples for Common Scenes

The widths below are assumptions chosen only to show the method. They are not measured data, and your scene will differ. Measure the width of your own scene where people will actually stand.

Example sceneAssumed widthRecognition: smallest resolutionIdentification: smallest resolution
Front door or porch10 ft381 px: 2MP / 1080p762 px: 2MP / 1080p
Garage door and driveway apron20 ft762 px: 2MP / 1080p1,524 px: 2MP / 1080p
Side yard or patio30 ft1,143 px: 2MP / 1080p2,286 px: 4MP
Small parking row60 ft2,286 px: 4MP4,572 px: More than 12MP
Wide yard or lot edge100 ft3,810 px: 8MP / 4K UHD7,620 px: More than 12MP

What the Tables Leave Out

  • Lens and distance. A pixel count says nothing about how wide the lens sees at a given distance. Use the lens distance calculator to turn focal length and sensor data into coverage.
  • Light and compression. Axis lists light direction, optics quality, and image compression as factors that affect the result.
  • The screen. Axis notes that when an external display is used, the ability to detect, observe, recognize, or identify depends highly on the screen’s resolution.
  • Storage and bandwidth. More pixels usually mean more data. Check the cost with the NVR storage calculator and the IP camera bandwidth calculator, and see how long cameras keep footage.
  • Other datasheet lines. Sensor size, IR, WDR, and lux matter as much as pixels. Our security camera specs explained guide covers them, and the NVR guide explains the recorder side.

FAQ

How many megapixels do I need for a security camera?

It depends on the scene width and the level of detail you need, not on a fixed megapixel number. Multiply the scene width in meters by the pixel density for your goal: 125 px/m for recognition or 250 px/m for identification under IEC 62676-4:2014, as summarized by Axis. The result is the horizontal pixels you need. A 20 ft wide scene needs 762 pixels for recognition and 1,524 for identification, and a 40 ft wide scene needs 1,524 and 3,048. Then pick the smallest resolution whose width in pixels is at least that number.

Is 2MP enough for a security camera?

Under the IEC 62676-4:2014 densities that Axis publishes, a 2MP / 1080p camera is 1,920 pixels wide. That covers a scene up to about 7.7 m (25 ft) wide for identification (250 px/m), 15.4 m (50 ft) for recognition (125 px/m), and 30.5 m (100 ft) for observation (63 px/m). Those are theoretical center-of-image values; lens quality, lighting, and compression can reduce usable detail, so leave a margin.

Do I need 4K (8MP) to identify someone?

Only if the scene is wide. An 8MP / 4K UHD camera is 3,840 pixels wide, which covers about 15.4 m (50 ft) at the identification density of 250 px/m. A scene about 25 ft wide or narrower is already covered by a 2MP camera at that density. Doubling the width in pixels from 1,920 to 3,840 is a fourfold increase in megapixels, and higher resolution generally needs more storage and bandwidth, which you can estimate with our NVR storage calculator and bandwidth calculator.

Is 5MP much better than 4MP for scene width?

Not for width. The 4MP size in our resolution comparison is 2,560 x 1,440 and the 5MP size is 2,592 x 1,944, so the 5MP camera is only 1.25% wider in pixels. The extra pixels are vertical. The scene width a camera can cover at a given density depends on horizontal pixels, so check the width in pixels on the datasheet.

Why do real cameras not match these tables?

The tables are a simplified model. Axis states that there is no guarantee that following the rule of thumb will let a camera fulfill the operational requirements, and that light direction, optics quality, and image compression also affect the result. Axis also notes that the specifications are valid when human operators interpret visual video, that analytics and thermal cameras use other definitions, and that the screen used to view the footage matters. Axis suggests that a higher pixel density can provide a safety margin for worst-case conditions.

What changed in IEC 62676-4:2025?

Axis documents the 2025 edition with a different set of levels, from Overview at 20 px/m to Scrutinize at 1500 px/m, including 125 px/m (Perceive) and 250 px/m (Characterize). This page uses the 2014 detection, observation, recognition, and identification densities, because those are the ones most datasheets and installers still quote. The calculator below accepts a custom px/m value if you want to test another level.

Sources

We calculated every figure on this page from the published densities and the pixel sizes listed in the sources. We did not test any camera.