Most quotes start with a number of cameras. That is the wrong end of the problem. A camera that shows "someone was here" sits in a different place to one that shows a face clearly enough to be useful later. Decide what each area has to deliver and the count works itself out.
Decide what the footage has to prove
For every area, pick one of three outcomes.
Observation. You want to know something happened and roughly how. Wide views over an aisle, car park or yard. Good for sequence of events, poor for identifying strangers.
Recognition. You want to confirm that the person is someone you already know — a staff member, a regular driver, a contractor. Less detail is needed because you are matching against a short list.
Identification. You want enough detail that someone who has never met the subject could pick them out. This is the demanding one, and it drives placement more than anything else.
A common way to express this is pixels per metre: horizontal pixels across one metre of scene at the distance that matters. Widely quoted starting figures look like this.
| Purpose | Rough pixels per metre | What you can expect |
|---|---|---|
| Detect that someone is there | 25–40 | A figure moving through the frame |
| Observe behaviour | 60–80 | Clothing, actions, direction |
| Recognise a known person | 100–130 | "That is our forklift driver" |
| Identify an unknown person | 200–250+ | A face worth giving to police |
Treat these as a starting point, not a standard: the same figure in daylight and in car park lighting at 2am are not the same result. Our resolution and lens guide works through how sensor size and lens choice change them.
Entry points first
Before worrying about coverage of open space, list every way a person or vehicle gets in or out:
- Front door and customer entrance
- Staff and back doors
- Roller doors and loading docks
- Gates, driveways and the boundary where the fence is weakest
- Internal doors to stock rooms, cash offices, server rooms and plant rooms
Each of those gets a camera positioned for identification: close, aimed across the opening rather than down the line of travel, and low enough that a face is in frame rather than the top of a cap. Entry cameras earn their keep because everyone who matters passes one, in both directions.
Only then add wide cameras for open areas: aisles, racking, the yard, the workshop floor, the till line.
Field of view, distance and overlap
Lens focal length decides how much you see and how far detail carries. On the sensor sizes common in business cameras, a rough guide is:
- 2.8mm — roughly 100 to 110 degrees across. Small rooms and the area around a door; detail falls away past about six to eight metres.
- 4mm — roughly 80 to 90 degrees. The workhorse for aisles and car parks, useful detail out to maybe twelve metres.
- 6mm — roughly 50 to 55 degrees, for longer driveways and aisles.
- 12mm and up — narrow views for a gate, dock door or plate camera.
Wide lenses do not create detail; they spread the same pixels over more scene. Two 4mm cameras covering a yard will usually beat one 2.8mm camera trying to cover all of it.
As a rough habit, let adjacent views share ten to twenty per cent of the scene, so a person leaving one frame appears in the next. Heavy overlap wastes cameras. Gaps are worse, because an insurer will find them.
Worked example: a small shop
A 120 square metre retail space with a front door, a back door through the stock room and a single counter.
- Front entrance, aimed at the door from inside, mounted about 2.4 metres — identification of everyone entering.
- Counter and till, looking across the transaction from the customer side so hands, goods and the register are all in frame.
- Sales floor wide view, covering the aisles and the blind side of the end-cap displays.
- Stock room and rear door, covering the doorway and the shelving nearest it.
Four cameras, with a fifth worth considering if deliveries are left in an external alcove. Two of the four exist to identify people, two to show behaviour.
Worked example: a warehouse
A 1,500 square metre warehouse with two roller doors, a pedestrian staff entry, an office and racking.
- One camera per roller door, aimed across the opening to catch faces and vehicle registrations as they pass through.
- Dock interior, showing the apron where goods are staged before loading. This is where most stock loss is proven or disproven, as our stock loss guide explains.
- Staff entry, for identification.
- Two to four aisle cameras, depending on racking layout, run down the aisle rather than across it.
- Office and stock cage doors.
- One external camera covering the hardstand and the gate approach.
Roughly eight to twelve cameras, moving with racking height and whether the yard needs its own coverage.
Worked example: an open yard
A fenced plant and equipment yard, 40 by 60 metres, with a gate and a site shed.
Open ground is deceptive. Pixels per metre at the far fence are a fraction of what they are at the gate, so yards need more cameras than owners expect, or narrower lenses.
- Gate camera, narrow lens, for vehicles and plates. The plate camera guide explains why that is a separate camera from the gate overview.
- Two corner cameras looking back along the fence lines, so a climber crosses the frame rather than walking towards it.
- One on the shed, one on fuel or high-value storage.
- A high wide camera for context, accepting it shows movement rather than identity.
Five to seven cameras, plus lighting. In a yard, lighting usually buys more than another camera does.
Count the cameras last
Work in this order: entry points, high-value areas, process areas, then context views. Add up what that list needs and you have a count you can defend. If the number is uncomfortable, cut areas rather than spreading cameras thinner — six cameras doing their job beat sixteen producing footage nobody can use.
Our camera count tool follows the same logic area by area, and the storage planner turns that count into retention days. For a plan on your own site, ask for a quote. Terms used here are in the glossary.