How do I calculate how many pixels I need at a doorway for facial identification quality?

IEC 62676-4 defines Identify level as 250 pixels per meter at the subject. An average adult face is approximately 0.2 meters wide, so you need 250 * 0.2 = 50 horizontal pixels across the face. A 1920-pixel-wide frame covering a 2-meter-wide doorway delivers 1920/2 = 960 ppm — well above the 250 ppm Identify threshold. The constraint becomes apparent when the camera is positioned to cover a wider area: a 1920p camera covering an 8-meter-wide lobby entrance delivers only 240 ppm, just below Identify threshold. The solution is either a dedicated face-capture camera aimed tightly at the 1–2 meter wide doorway, or a higher-resolution (4K = 3840 horizontal pixels) camera that can cover the wider area while maintaining 250+ ppm.

What focal length should I use for a camera covering a parking lot entrance at 15 meters?

Work backward from your pixel density requirement. For vehicle overview (Observe level, 62 ppm) of a 6-meter-wide entrance lane at 15 meters with a 1/2.8-inch sensor 1080p camera: required coverage width = 6 m, delivering 1920/6 = 320 ppm — Identify level, more than sufficient. The focal length for 6 m coverage at 15 m: using a 1/2.8-inch sensor (5.4 mm width), required FoV = 2*arctan(3/15) = 22.6°, which corresponds to approximately a 13–14 mm focal length lens. For license plate capture you want a separate dedicated LPR camera with a longer focal length (25–50 mm), higher shutter speed, and IR illumination aimed along the lane rather than across it.

Is a single 4K fisheye camera a cost-effective replacement for multiple conventional cameras in a large open room?

For Detect and Observe level requirements in an open room without obstructions, yes — a single 8MP fisheye camera at ceiling center can provide panoramic coverage of rooms up to 15x15 meters with acceptable pixel density for scene monitoring and general tracking. The critical limitation is pixel density at the room perimeter: in a 12-meter diameter room, the dewarped image from an 8MP fisheye camera delivers approximately 60–80 ppm at 6 meters from the camera — marginal for Observe, insufficient for Recognize or Identify. Do not use fisheye cameras as the sole coverage for positions where facial identification may be required. The cost-effective hybrid design uses one fisheye for room overview context plus two to four fixed cameras at entry/exit points for Identify-level forensic coverage.

What is variable bitrate (VBR) and how much storage does it actually save?

Variable bitrate (VBR) encoding adjusts the compression ratio frame-by-frame based on scene complexity. A static parking lot at 3 AM with no movement can be compressed at 0.5–1 Mbps (high compression ratio on redundant background frames) without visible quality loss. The same camera during a busy daytime scene with vehicle movement encodes at 4–8 Mbps to preserve motion detail. VBR with H.265 and motion-triggered quality boost (sometimes called "smart codec" or "smart stream" by manufacturers) typically reduces average bitrate by 40–70% compared to constant bitrate (CBR) settings in typical security applications with mixed busy/quiet periods. Over a 30-day retention period, this translates directly into proportional storage reduction — a system designed for 130 TB CBR may require only 50–70 TB in practice with VBR enabled.

How much camera field of view overlap is needed between adjacent cameras to avoid coverage gaps?

Best practice is 15–20% horizontal FoV overlap between adjacent cameras at the maximum required DORI coverage distance. This means at the far edge of each camera's specified coverage range, the adjacent camera has already picked up the scene with 15–20% of its own frame. Below 10% overlap, a subject moving between cameras will have a brief coverage gap during which their position and actions are unrecorded — a gap that defense attorneys regularly exploit in prosecutions. Above 30% overlap, you are over-specifying camera density and adding unnecessary hardware cost. When designing layouts in JVSG or similar tools, set your FoV footprints to overlap by 15–20% at the specified DORI range, not at the camera's maximum physical detection distance.

← Physical Security Studio

Engineering·10 min read·July 1, 2025

🔒 Security Camera Layout and Coverage Analysis: Field of View, Mounting Height, and Lens Selection

Engineering methodology for designing security camera coverage — calculating field of view, selecting focal lengths, determining mounting heights, performing coverage gap analysis, and documenting camera layouts to IEC 62676-4 pixel density requirements.

The DORI Framework: Detect, Observe, Recognize, Identify

IEC 62676-4 (Video Surveillance Systems for Use in Security Applications — Part 4: Application Guidelines) defines the DORI framework as the standard methodology for specifying camera performance requirements. DORI stands for four image quality levels, each defined by a minimum pixel density at the subject:

Detect (D) — 25 pixels per meter (ppm). The minimum level for detecting that a human or vehicle is present in the scene. Adequate for open area monitoring and parking lot overview.
Observe (O) — 62 ppm. Sufficient to observe general clothing, body posture, and movements of a subject. Used for corridor and lobby overview cameras.
Recognize (R) — 125 ppm. Enables confirmation that an observed person matches a known individual from a reference image. Required for secondary exit/entry confirmation.
Identify (I) — 250 ppm. Forensic-quality imagery enabling reliable identification of an unknown subject from camera footage alone. Required at primary entry points, ATMs, cashier positions, and evidence-quality coverage zones.

The DORI thresholds assume adequate illumination (minimum 30 lux for color, 10 lux for monochrome at the subject), proper focus, and a stationary or slow-moving subject. Specify the required DORI level for each camera position before selecting hardware — many projects under-specify and end up with inadequate forensic coverage at the most critical points.

Field of View Calculation: The Core Design Math

The horizontal field of view (FoV) of a camera depends on the sensor size and focal length according to the fundamental optics equation:

FoV (degrees) = 2 * arctan(sensor width / (2 * focal length))

For a common 1/2.8-inch sensor (horizontal sensor width approximately 5.4 mm) and a 4 mm focal length lens: FoV = 2 * arctan(5.4 / 8) = 2 * 34° = 68° horizontal FoV.

The horizontal coverage width at a given distance is: Width (m) = 2 * distance * tan(FoV/2). For the above camera at 10 meters: Width = 2 * 10 * tan(34°) = 13.6 meters.

To achieve a pixel density requirement: minimum camera resolution (pixels) = required PPM * coverage width. For 1920px horizontal resolution covering 13.6 m: PPM = 1920/13.6 = 141 ppm — sufficient for the Recognize threshold at 10 meters. Increase focal length (narrower FoV) to increase PPM at a given distance; decrease focal length (wider FoV) to cover more area at the cost of pixel density.

Mounting Height: Balancing Coverage and Face Capture Angle

Mounting height directly affects two competing requirements: coverage area (higher mounting = wider field, more area covered per camera) and face/subject identification quality (lower mounting + downward angle = better facial geometry capture). The optimal mounting height depends on the DORI requirement for the zone:

Detect/Observe zones (lobbies, parking lots) — mount at 3–5 meters. Wide area coverage, overhead angle is acceptable for general scene monitoring. Pan-tilt coverage can extend effective range to 15–30 meters at Detect level.
Identify zones (doorways, checkout counters, ATMs) — mount at 2.0–2.5 meters, aimed to capture the face at a downward angle of 15–30 degrees from horizontal. Above 30 degrees of downward angle, the facial geometry becomes distorted and recognition accuracy degrades significantly.
Vehicle license plate capture — dedicated license plate recognition (LPR) cameras should be mounted at bumper-to-bumper height (0.5–0.8 m above grade), angled 5–15 degrees downward, and aligned with the vehicle travel path. High-shutter-speed lenses (1/4000s minimum) and supplemental infrared illumination are required for moving vehicles.

At entry chokepoints, use two cameras: a wide-angle overview camera at 3–4 meter height for scene context, and a dedicated face-capture camera at 2.0–2.2 meters aimed directly at the entry point. This two-camera pairing strategy is industry best practice for forensic-quality evidence at controlled access points.

Lens Selection: Fixed vs. Varifocal vs. Motorized Zoom

Fixed focal length lenses — pre-set optics (2.8 mm, 4 mm, 6 mm, 8 mm, 12 mm are common in IP camera product lines). Lowest cost, highest optical quality for a given price point, no moving parts. Select for positions where the camera location, mounting height, and required coverage width are precisely known and will not change.
Varifocal lenses (manual zoom) — adjustable focal length set at installation (typically 2.8–12 mm or 6–22 mm ranges). The most common choice for general security cameras because field adjustment corrects site design errors and accommodates future layout changes. More expensive than fixed, slightly lower optical performance at extreme zoom settings.
Motorized/auto-focus varifocal — electrically adjusted focal length and focus via camera menu. Essential for cameras installed in difficult-access locations (ceiling plenums, high mounts, building exteriors) where manual lens adjustment would require a lift or scaffolding. Most modern IP cameras in the mid-range and above include motorized varifocal as standard.
Fisheye/360-degree lenses — ultra-wide angle (180°+ FoV) using curved lens elements. A single fisheye camera at ceiling center of a room can replace 3–4 conventional cameras for room overview, but pixel density at the perimeter of the dewarped image is significantly lower than at the center. Use only for Detect/Observe requirements in interior spaces, never as a primary Identify-level camera.

Coverage Gap Analysis and the Camera Layout Drawing

A professional camera layout design produces a scaled floor plan or site plan with each camera's field of view drawn at the specified DORI distance. Gap analysis identifies areas with no camera coverage, areas with only Detect-level coverage where Identify is required, and blind spots created by architectural obstructions (columns, stairwells, elevator shafts).

Overlap planning: 15–20% horizontal FoV overlap between adjacent cameras at the required DORI range ensures that a subject moving through the scene is visible to at least two cameras simultaneously. This overlap zone is critical for maintaining continuous tracking in a scene-to-scene handoff workflow and for providing redundant evidence if one camera is blocked or fails.

Tools: Jvsg IP Video System Design Tool and CCTV Design Lens Calculator are industry-standard software for camera layout, FoV projection, and coverage documentation. Output should include a camera schedule (camera number, model, focal length, mounting height, DORI requirement, DORI achieved) and a site plan with FoV footprints drawn to scale.

NVR/VMS Bandwidth and Storage Dimensioning

After determining camera count, resolution, and positions, engineers must dimension the recording infrastructure. Key parameters:

Bitrate per camera — H.265 4K camera: 8–16 Mbps continuous; H.265 1080p: 3–6 Mbps. Enable variable bitrate (VBR) with motion-triggered high-quality mode to reduce average storage consumption by 40–60% in low-activity scenes.
Retention period — specify minimum retention by regulatory requirement. TSA mandates 90-day retention for airport sterile area cameras; NISPOM requires 90 days for government cleared facilities. Many enterprise security policies require 30 days minimum, with 90 days for high-security areas.
Storage calculation — Total storage (TB) = (cameras * avg bitrate Mbps * 0.125 MB/Mb * 3600 s/hr * 24 hr * retention days) / 1000. For 100 cameras at 4 Mbps average, 30-day retention: (100 * 4 * 0.125 * 86400 * 30) / 1000 = 129.6 TB raw. Add 20% overhead for RAID and OS. Plan for RAID 6 at minimum for NVR storage arrays in security applications.
Network infrastructure — PoE switch budget: 1920p cameras draw 8–13W per port (IEEE 802.3at PoE+). 4K cameras with PTZ heaters may draw up to 25.5W (IEEE 802.3bt PoE++). Aggregate switch uplinks to NVR must carry sum of all camera bitrates with 30% headroom for motion spikes.

Topics covered

security camera layoutCCTV field of viewcamera mounting heightlens focal length selectionIEC 62676-4pixel density CCTVcamera coverage analysisCCTV designfisheye camera securityvarifocal lens securityPPM pixels per meterDORI standardsecurity camera placement guidecamera coverage overlapNVR camera resolution

🛠️ Related Free Tools

Put this knowledge to work on your iPhone

Browse our full catalog of professional iOS apps — from electrical code tools to AI builders.

Browse All 95+ Apps