CCTV Camera Selection and Placement

The most common fault in a CCTV system is not a broken camera. It is a working camera producing footage that cannot be used for the thing it was installed to do. Someone reviews an incident, sees a person clearly enough to know a person was there, and cannot say who. That outcome is decided at the design stage, by the lens and the mounting position, not by the recorder.

Start with the task, not the camera

Before any equipment is chosen, each area needs a stated purpose. Do you need to know that someone is present, or to recognise a person you already know, or to identify a stranger from the footage alone? These are different requirements with very different costs, and a site rarely needs the highest one everywhere.

A car park entrance might need identification at the gate and only detection across the rest of the tarmac. Specifying identification across the whole area is how budgets get spent on cameras nobody needed.

The DORI framework

EN 62676-4, the application guidelines part of the European video surveillance standard, turns those tasks into numbers. It anchors four named tasks to four pixel densities measured across the target, expressed in pixels per metre:

  • Detection at 25 px/m - is a person or vehicle present
  • Observation at 62.5 px/m - characteristic details such as clothing, and the activity around an incident
  • Recognition at 125 px/m - a high degree of certainty that this is someone previously seen
  • Identification at 250 px/m - identifying an individual from the image alone

Because pixel density falls off with distance, one camera delivers all four at different ranges. A camera rated for identification at 10 metres will only manage detection at around 100 metres. That relationship is the single most useful thing to understand when reading a datasheet.

These thresholds are what EU public tenders and insurance auditors reference, so a design that states its DORI targets per area is far easier to defend than one that lists camera models.

The 2025 amendment

The 2025 edition of IEC 62676-4 replaces DORI with a longer seven-step ladder, on the basis that 250 px/m was not always enough to identify people in low light or where the subject is moving. The identification-equivalent threshold rises to 500 px/m and the zone is renamed Validation.

The older DORI ladder is still what most specifications and tender documents cite, so both are worth knowing. The practical reading is that 250 px/m was always a floor rather than a comfortable target, and designing with headroom above it was never wasted effort.

Lens, sensor and field of view

Pixel density is a function of the sensor resolution, the lens focal length and the distance to the target. A wider field of view spreads the same pixels across more metres, so every increase in coverage costs detail. Two cameras with narrower views will usually beat one wide camera trying to do both jobs.

Varifocal lenses give room to adjust on site, which matters because the scene is never quite what the plan assumed. Fixed lenses are cheaper and perfectly adequate where the geometry is known and will not change.

Mounting and lighting

Mounting height is a trade-off. Too low and the camera is reachable and easily defeated. Too high and you are looking at the tops of heads, which is useless for identification. For faces, the camera wants to be near head height at the point of capture, which is why entrance cameras are often mounted lower than instinct suggests.

Lighting decides more than most people expect. A camera pointed at a glass entrance with daylight behind it will silhouette everyone walking through. Wide dynamic range helps but does not eliminate the problem, and repositioning to avoid shooting into the light is usually the better fix. For night coverage, infrared illumination has a real working range that is often shorter than the marketing figure.

Training

The CCTV and Video Surveillance Fundamentals course covers camera selection, lens calculation and system design in structured form. The UK equivalent is the CCTV Installation Course.

Once cameras are chosen, the next questions are recording and retention - see CCTV storage, retention and data protection.