CCTV Installation Best Practices for Office, Warehouse and Production Areas

1. What "best practice" means for CCTV

Best practice in CCTV installation is not a marketing claim — it is a measurable standard. In Europe, the reference document is EN 62676-4 "Video surveillance systems for use in security applications — Part 4: Application guidelines". EN 62676-4 defines the four operational tasks a camera can serve (Detect, Observe, Recognize, Identify), assigns each a minimum pixel-density target on the subject, and requires the design package to prove those numbers are met on the floor plan and on the wall after install.

The four DORI categories are: Detect 25 px/m (a person is present), Observe 62 px/m (read what they are doing), Recognize 125 px/m (match against a known person), and Identify 250 px/m (legally usable identification of a stranger). Every camera you install must be tied to one of these targets at a documented distance — not just "looks good on the monitor". Use the pixel-density (OODPCVS) calculator to verify each zone before quoting the job.

The second baseline is legal. GDPR Article 13 (and its Polish transposition under RODO) requires that anyone entering a monitored area is informed before the recording starts. That means signage at every entrance, a documented retention period (30 days is the EU default unless a longer period is justified), a named controller, and a privacy notice the subject can read. A technically perfect install that lacks signage is unlawful and the recordings cannot be used as evidence — courts will throw them out.

Best practice therefore has two halves: technical (EN 62676-4 DORI, IP and NEMA ratings, PoE budgets, NTP-stamped recordings) and legal (GDPR signage, retention, controller documentation, lawful basis). Skipping either half voids the value of the system. The rest of this guide assumes you intend to deliver both.

2. Site survey checklist before any cable is pulled

The site survey is where margin is won or lost on a CCTV job. Twenty minutes with a tape, a lux meter and a phone camera will save twenty hours of rework. Show up with a printed floor plan, a clipboard and a coloured marker. Do not rely on memory or photos alone.

3. Mounting heights by environment

Mounting height drives everything downstream: lens choice, tilt angle, blind-cone size and the realistic DORI distance. The numbers below are working ranges that most integrators converge on, not absolutes.

The principal trade-off is reach versus near-field coverage. A camera at 8 m looking 30° down has a substantial blind cone directly beneath it — typically 2 m wide at floor level — and a sharp drop in facial pixel density at any subject closer than about 4 m horizontal. The fix is not to mount lower; the fix is to add a second camera that covers the near field of the first. This is "cross-coverage" and it is non-negotiable for high-value zones.

Use the lens-selector calculator to pick a lens that puts your DORI target inside the camera's effective working range for the chosen height. The same calculator flags when the height you have proposed is incompatible with your DORI requirement — for example, a 12 m mount cannot deliver Identify-level 250 px/m at 20 m on a 4 MP sensor.

4. Office areas: reception, open plan, server room, meeting rooms

Office CCTV is almost always Identify-level at the perimeter and Observe-level inside. Privacy expectations are higher than in warehouses, so over-coverage attracts works-council complaints and GDPR audits. Cover what you must, not what you can.

Reception and main entrance: one camera looking down the approach toward the door at 2.5 m on a 6 mm lens for Identify (250 px/m) at the threshold, plus one wide 2.8 mm dome inside reception for context. Avoid pointing the entrance camera straight out through glass — solar back-light will blind it. Use a camera with true WDR (120 dB+) or move the camera off-axis.

Open-plan workspace: consult the works council before you specify anything. Where coverage is justified, mount 2.8-4 mm domes at 2.7-3.0 m at intersections of partitions, aimed to cover circulation routes rather than individual desks. Desks themselves should not be in the centre of FOV.

Server room / comms room: one camera covering the door from inside, mounted at 2.4 m on a 2.8 mm lens. The aim is to capture every person entering. A second camera covering the cold aisle is useful for audit but optional. Server-room CCTV is the easiest zone to justify under "asset protection" lawful basis.

Meeting rooms: generally do not install CCTV inside meeting rooms. Confidential conversations and works-council rules make it indefensible. Cover the corridor outside instead.

5. Warehouse zones: docks, aisles, racking, staging, perimeter

Warehouse CCTV is the bread and butter of professional installers. Each zone has its own DORI target and its own lens story. Get this wrong and you end up with a 60-camera install that cannot identify the operator who damaged the rack. See the dedicated warehouse CCTV deep-dive for full layouts, and the warehouse use case for sample bills of quantities.

Loading docks: Identify-level (250 px/m) on faces at the dock door and Recognize-level (125 px/m) on goods. One 4 MP camera at 4 m on a 4-6 mm lens, aimed along the truck approach, will deliver Identify out to about 8 m. Add a second camera inside the dock door looking back outwards to capture the driver as they exit the cab.

Aisles between racking: mount on the rack itself, not the ceiling. A camera at the end of each aisle at 5-6 m on a 4 mm lens, looking down the aisle, gives Recognize-level coverage of the whole aisle. Ceiling-mounted cameras across racks see only the top of the topmost pallet — useless. Verify your FOV cones on the floor plan against actual rack heights.

Staging and pick areas: Observe-level (62 px/m). Wide 2.8 mm domes at 5-6 m, spaced for 15-20 percent FOV overlap. These zones see the highest staff density and the most product movement, so coverage continuity matters more than identify-level detail.

Perimeter and yard: bullets, not domes. Mount at 4-5 m on 6-12 mm varifocal lenses, aimed parallel to the fence line so an intruder must cross multiple FOVs. Confirm DORI distance for each camera against the actual fence-to-camera offset.

6. Production areas: machinery, hazmat, IP ratings, vibration mounts

Production floors are the most hostile environment for CCTV. Dust, oil mist, washdown water, vibration from presses and conveyors, EMI from inverters, and temperature swings from 5°C in unheated bays to 60°C above ovens. Spec a camera that is wrong for the environment and it will fail inside 12 months, often catastrophically.

IP ratings — what each digit means. The first digit is dust ingress (0-6, where 6 is dust-tight). The second is water ingress (0-9). The combinations you will actually quote:

NEMA 4X for chemical environments. IP ratings cover physical ingress but not chemical attack. NEMA 4X (typically 316L stainless housing) adds corrosion resistance against salt spray, acids, alkalis and food-grade cleaning chemicals. Specify NEMA 4X for coastal sites, de-icing-salt-exposed loading yards, chemical plants and any food production line using CIP (clean-in-place) chemistry. Powder-coated aluminium housings will corrode through within 18-24 months in these conditions.

Vibration isolation mounts. Cameras mounted directly to press frames, conveyor structures or stamping lines will lose focus, suffer bearing wear in the IR cut filter motor, and ultimately delaminate the sensor bond. Use a vibration-isolating mount (rubber-isolated bracket, minimum 20 mm elastomer pad) for any camera within 3 m of equipment producing more than 0.5 g RMS vibration. For very high vibration, mount the camera on a separate freestanding pole or wall and accept the longer cable run.

Hazmat and ATEX zones. Any zone classified ATEX (potentially explosive atmosphere — flour mills, paint booths, fuel storage) requires a certified explosion-proof housing, typically stainless 316L with flame paths cast into the lid. Do not improvise this. The certification cost is high but the legal liability for a non-certified camera in an ATEX zone is far higher.

7. Cable runs, PoE budget, conduit and EMI separation

Cabling is where most installs accumulate hidden technical debt. The cameras might be perfect, but a 110 m run on Cat6 with EMI from a nearby VFD will give you intermittent dropouts that no one connects back to the cable for months.

PoE standards and budgets. PoE 802.3af (PoE Type 1) delivers 15.4 W at the source, ~12.95 W at the device — enough for a fixed dome. PoE+ 802.3at (PoE Type 2) delivers 30 W at source / ~25.5 W at device — enough for PTZ, IR-heavy bullets and most heated outdoor cameras. PoE++ 802.3bt Type 3 delivers 60 W (~51 W at device); Type 4 delivers 90 W (~71 W at device) — needed for high-power heaters, ePTZ multi-sensors and integrated illuminators.

When you sum the PoE budget, check the switch's total PoE budget, not just the per-port maximum. A 24-port switch rated PoE+ per port often has a total budget of only 370 W — meaning you cannot deliver 30 W on every port simultaneously. If you have 24 bullets at 25 W each, you need 600 W of switch budget, which means a 740 W or 1000 W class.

Cat6 vs Cat6A. Cat6 is enough for 1 Gbps PoE+ runs up to 55 m. Beyond 55 m, or anywhere you might want to upgrade to 4K cameras and PoE++ later, use Cat6A — it is rated for 10 Gbps to 100 m and handles the heat dissipation of PoE++ better. The 100 m PoE limit is hard: it is the IEEE limit on Ethernet, not a guideline. For runs over 100 m use fiber with a media converter at each end, or a mid-span PoE injector that re-clocks the signal.

Conduit and EMI separation. Run CCTV cable in its own conduit, not shared with mains or motor lines. Maintain 300 mm minimum parallel separation from 230 V / 400 V power, 500 mm from VFDs and motor feeds, and cross power lines at 90° where they must intersect. In heavy-industrial environments (welding shops, large motor halls), use shielded Cat6A S/FTP and ground the shield at one end only — grounding both ends creates a ground loop that injects 50 Hz hum into the video.

Fire compartments. Every cable crossing a fire boundary needs an approved intumescent collar or putty seal. Audit-proof the install by photographing each penetration with a tape measure showing the seal thickness. Building control will require this and so will the insurer after any claim.

8. Lens and resolution per zone

Lens choice is the single most consequential decision after camera position. A 4 MP camera with the wrong lens delivers less usable evidence than a 2 MP camera with the right one. Use the FOV calculator and DORI calculator together — the first tells you what the camera sees, the second tells you whether the sight is legally useful.

Resolution interacts with lens. A 2 MP camera on a 2.8 mm lens at 3 m mounting height delivers ~120 px/m at 5 m horizontal — Observe, just short of Recognize. The same camera at 8 MP delivers ~240 px/m at the same distance — almost Identify. Doubling resolution is usually cheaper than doubling camera count, but only if your network and storage are sized for it.

Use the lens selector to match each camera position against its DORI target. The selector will flag combinations where the lens is too narrow (gaps at the sides) or too wide (insufficient pixel density at the working distance). For vertical-specific layouts see the retail and parking use cases.

9. Lighting, glare, IR bleed and back-light compensation

A CCTV camera is a light sensor with a lens in front of it. If the light is wrong, no camera setting will rescue the image. The most common lighting failures are back-light, IR bleed and IR hotspots.

WDR (Wide Dynamic Range). WDR cameras combine multiple exposures into a single frame, so bright windows and dark interiors are both captured. Real WDR (sensor-level, sometimes called "true WDR") is 120 dB or higher. "Digital WDR" or "DWDR" is software contrast stretch and does not actually capture more dynamic range — avoid it for any entrance facing daylight. Quote the dB number, not the marketing label.

BLC (Back-Light Compensation). BLC darkens the bright area and lifts the dark area in a fixed region of interest. It is older than WDR and produces a more contrasted result. Useful where there is a single dominant back-light source (an entrance door facing south) and the subject is always in the same place. Set the BLC zone on the subject area, not the background.

HLC (High-Light Compensation). HLC selectively masks bright point sources — vehicle headlights, sunlight reflections — so the rest of the frame stays exposed. Essential for parking entrances and night-time perimeters where headlights would otherwise blow out plate detail.

IR bleed. Built-in IR LEDs sit a few centimetres from the lens. In rain, fog or dust, IR light scatters off particles directly into the lens, producing a foggy white image. Either disable on-board IR and use a separate IR illuminator mounted 1-2 m away from the camera, or specify a camera with a long IR baffle. For exterior cameras under eaves with spider webs, IR bleed is the number-one false-alarm trigger on motion analytics.

IR hotspots. A bright white object close to the camera (a numberplate, a high-vis vest, a reflective floor stripe) will reflect IR back into the sensor and overexpose. Verify night performance by walking a person in high-vis through the FOV during commissioning, not just by looking at the empty scene.

10. Commissioning: focus, NTP, retention, GDPR signage

Commissioning is where the install is signed off. A rushed commission produces a system that looks fine on day one and fails the first time it is needed. Allocate a full day for a 20-camera site and follow a checklist.

Focus calibration. Auto-focus is unreliable at night and in IR mode. Manually focus each camera using a focus chart placed at the planned DORI distance, with the iris fully open (to expose focus errors), then lock the focus ring. Re-check focus in IR mode — back-focus shifts between visible and IR for most fixed lenses, and a camera focused in daylight may be soft at night.

NTP for legal timestamps. Configure every camera and the NVR to a single NTP source — ideally the on-site domain controller or a hardware GPS clock, not a public NTP pool that can drift or be unreachable. Without synchronised NTP, recordings carry timestamps that cannot be cross-referenced against access logs, alarm logs or other CCTV systems, and a defence lawyer will use that to exclude the evidence.

Retention period. The EU default is 30 days unless a longer period is justified by purpose. Set retention per zone if your NVR supports it — perimeter at 30 days, server room at 90 days, public areas at 7-14 days. Document the retention decision in the system manual; the data protection authority will ask for it.

GDPR signage. Post Article 13 notices at every entrance before commissioning ends. The notice must include: the fact of CCTV recording, the controller's name and contact, the purpose, the retention period, the lawful basis (usually legitimate interest), and a pointer to where the full privacy notice is available. A photo of each posted notice goes in the as-built handover pack.

As-built documentation. Hand over a single PDF containing: floor plan with camera positions, IDs and FOV cones; PoE budget sheet per switch port; IP and lens spec per camera; retention configuration; NTP source; signage photos; and a maintenance schedule. Use the planner export from CCTVplanner as the floor plan source so the as-built matches the design.

11. Common installation mistakes and how to audit them

These seven mistakes account for the overwhelming majority of CCTV installs that have to be redone within the first two years. Audit your design and your competitor's design against this list before the camera goes on the wall.

For terminology used in this guide — DORI, PoE classes, IP/NEMA, WDR, BLC, HLC — see the glossary.