How Many Cameras Does It Take to Cover One Hectare?

The short answer: two to four cameras per hectare — and the hectare is not what decides it

One hectare is 10,000 m², a square roughly 100 m on a side. Most working sites that size are covered properly by two to four cameras, and a five-hectare farm usually runs on six to ten, not the fifteen or twenty a straight area calculation suggests. Camera counts do not scale with area, because nobody watches ground. You watch assets and access points: the pump house, the equipment yard, the fuel drum, the one gate a truck has to come through. A hectare of open pasture in the middle of a block needs zero cameras. A hectare with a workshop, a gate and a diesel tank on it needs three.

Plan the count from a map, not from a hectare figure. Here is the method, the geometry behind it, and what changes on a fenced yard or a long thin site like a solar array.

Why "area coverage" is the wrong unit

A camera does not cover an area. It resolves detail out to a distance, and the amount of detail collapses as that distance grows: pixels on target fall off roughly in proportion to range, so a face that is identifiable at the gate is an unrecognisable smear at the far fence.

Detect, recognise, identify: three different distances

The industry way of describing this is the DORI scale in IEC 62676-4 — Detect, Observe, Recognise, Identify. The same camera gives you four completely different working distances depending on which of those four jobs you need. Ask any supplier, including us, for the DORI table of the exact SKU you are quoting rather than a headline "range" number, because a headline range always quotes the easiest of the four.

That gives you a two-tier plan instead of a coverage plan:

Trying to buy identification-grade coverage across a whole hectare is what makes quotes absurd. You would need dozens of cameras. Detection across the open ground plus identification at the three or four pinch points is what a real site actually buys.

The geometry that matters: perimeter, not area

This is the number that surprises buyers. Double the area and you do not double the work, because perimeter grows with the square root of area while area grows linearly.

A one-hectare block is the worst ratio you will ever buy at: it is almost all edge. That is why per-hectare camera density falls sharply as sites get bigger, and why quoting "X cameras per hectare" to a 50-hectare client will lose you the job on price for no security benefit.

A ten-minute counting method

Print the satellite image. Then:

1. Mark every asset worth stealing. Machinery, fuel, pumps, wire, batteries, irrigation controllers, stored crop, livestock handling pens. Cluster anything within about 20 m of another item — one camera usually takes a cluster.
2. Mark every way in. Gates, cattle grids, the gap in the hedge everyone actually uses, the neighbour's track, the riverbank. Access points are where identification happens.
3. Draw the sightlines. From each candidate mounting point, is there an unobstructed view to the asset or the gate? Trees, silos, container stacks and standing crop kill more plans than distance does.
4. Assign one camera per pinch point, one per asset cluster, one per blind corner. Add nothing else on the first pass.
5. Only then look at the fence line and ask whether detection along the open edges is genuinely worth the extra units, or whether the gates already catch anyone who has to leave with something heavy.

Typical counts by site type

Linear sites break the rule

A hectare shaped as a 20 m × 500 m strip — a canal bank, a plantation access road, a pipeline easement, a row of solar tables — has a 1,040 m perimeter for the same 10,000 m². Nothing about the hectare figure predicts the camera count here.

For linear sites, stop counting hectares and count per kilometre of run and per structure. Cameras look along the line rather than across it, so each unit earns much more distance than it would in an open square. The units cluster at the ends, at road crossings, and at the equipment that is actually worth stealing — on a solar site that is the inverter and transformer stations and the delivery gate, not the middle of a row.

Why off-grid units usually cut the count

Cabled systems distort camera counts in a way most buyers never notice. Because every unit has to reach an NVR and a power supply, cameras end up clustered where the trenching already exists, and you then buy extra cameras to see the far corner by overlapping runs from the wrong place. The trench, not the threat, sets the layout.

A solar-plus-4G unit has no trench and no NVR run. It goes on the pole nearest the thing you care about, and it goes up in 5–10 minutes. The third camera does not cost four times the second one simply because it sits 180 m past the last junction box. In practice that means fewer cameras doing more useful work — and it is why a 5 ha farm quote comes in under the cabled equivalent even before the civils are priced.

Practical consequences for the plan:

• Mounting points are free choices, so put the identification camera where a face is unavoidable, not where the conduit ends.
• Sun and signal replace cable as the constraints. A panel shaded by a mango tree at 09:00 in the wet season is a support ticket in month three, and the far corners need bars from the carrier you are actually putting SIMs on, not the one with the prettiest coverage map.

Our catalogue and specification tables are at /en/products/, including which models are fixed lens and which are PTZ — the trade-off between wide detection and zooming into a pinch point changes the count too.

The failure that no camera count fixes: cameras that are asleep

The most common reason a farm buys the "right" number of cameras and still has nothing to show the police is that the cameras were not recording. PIR-triggered units sleep until infrared movement crosses a sensor whose working distance is far shorter than what the lens can see, so a person walking at the edge of frame, a slow approach along a hedge, or a vehicle drifting in with the engine off never wakes it. You get a clip that starts after the event and ends before the plate is visible.

That matters more on farmland than anywhere else because of when it happens. Theft happens between roughly 01:00 and 04:00, on public holidays, during harvest when everyone is exhausted, and in the weeks when a block is between crops and nobody drives past. Those are precisely the hours nobody is checking the app.

Where AOV changes the odds

AOV (Always-On Video) is the industry-generic answer: continuous recording at low standby power, with AI scene analysis deciding what to flag rather than a passive sensor deciding what to record. The count on your map does not change — the odds that those cameras have the incident on file do. The full technical case, including what it costs in power and storage, is in our AOV guide.

What to put in the quote

Five things to confirm with the customer before you send a number:

On the commercial side: the catalogue runs to 150 SKUs, with wholesale prices quoted per order, all available for OEM or private label. Outdoor models are IP66, resolution is 1080p–2K as standard with 4K on custom or OEM builds, and storage is microSD up to 128 GB, larger only on custom orders. The manufacturer is Shenzhen Leksell, established 2012, certified by TÜV Rheinland with CE, FCC and RoHS.

Trial orders start at 10 units — two or three farm installs plus a demo — and OEM/ODM at 100, with dealer margin at 20% from around 10 units and 30% from around 50. Samples ship in 7–10 days by air, bulk in 25–40 days by sea, and stock into Vietnam through our Ho Chi Minh City office in 7–10 business days.

Quoting a specific block? Send the satellite image with the assets marked to info@solaguard.net or WhatsApp +86 188 1869 9870 and we will mark up a count against it. Volume tiers are at /en/wholesale/.