Practice
Which camera for a driveway, a front door and a warehouse
2 min readcctvfinder
The question always arrives in the same shape: "which camera do I need for my driveway?" There is no answer to that, because it depends on four numbers. Here are five typical spots, all four numbers stated, and each one ready to reproduce.
Driveway, reading a number plate
Twelve metres to the barrier, camera at three metres, a 22-degree lens, 4K. Result: 818 pixels per metre. That is well above the 250 that Identify asks for.
The catch is in the width: 4.7 metres. That is one traffic lane and nothing else. Anyone wanting the same camera to cover the footpath beside it loses the number plate again.
Number plate capture is its own camera with its own job. It almost never fails on resolution.
Driveway, a person at the gate
The same distance, but now it should be recognisable who is standing there. Twelve metres, three metres high, 45 degrees, Full HD. Result: 192 pixels per metre across ten metres of width.
That is enough for Recognise with room to spare, and it comes from a camera in the lower price bracket. Ten metres of width covers the gate and its surroundings. Two jobs, two cameras, and neither is spectacular.
Front door, a face at the threshold
Five metres, camera at 2.8 metres, 50 degrees, Full HD. Result: 400 pixels per metre across 4.8 metres of width.
At short range Identify is nearly free. The critical value at a front door is therefore not the resolution but the mounting height: hang the camera too high and it looks at the crown of the head and the shoulders instead of the face.
Warehouse, an aisle lengthways
A 25-metre aisle, camera at six metres below the ceiling, 40 degrees, 4K. Result: 208 pixels per metre across 18.5 metres of width.
The aisle is covered to Recognise along its whole length. With a 90-degree lens on the same camera it would be 76 pixels per metre, so only Observe. Same hardware, two quality levels apart.
Warehouse, the loading door from inside
Fifteen metres to the door, camera at five metres, 50 degrees, Full HD. Result: 134 pixels per metre across 14.3 metres of width.
That sits just above the requirement for Recognise. Just above is a warning at this spot, not a result: at night nothing is left of it, because a sensor short of light either amplifies or exposes longer.
Plan a spot to the threshold and you have planned it for sunshine.
All five to reproduce
Each link opens the 3D view with exactly these values. From there you can keep turning: another resolution, another angle, another distance.
Read next
Calculating field of view: which angle covers how much width
Two lines of arithmetic and a table: how wide a camera sees at 10, 20, 30 and 50 metres, and how many pixels per metre are left.
2 minTechnologyNight vision and IR range: what the data sheet figure really says
A data sheet promises 50 metres of infrared range. What was measured is something other than what you want to see at night.
2 min