9 min read
Start with the question
Camera count is the wrong starting point. The right starting question is: what events does this facility need to see, at what level of detail, in what lighting, and for how long should the footage be retained? Every answer drives a different camera decision — sometimes very different.
Coverage goals
- Monitor — general activity awareness in an area
- Detect — see that something has happened
- Observe — track a person's general appearance and behavior
- Recognize — identify a known person
- Identify — produce footage that supports identification of an unknown person
Pixel density and the level you need
Each coverage goal corresponds to a different pixel density at the scene. Monitoring takes very little; identification takes a lot. Designing every camera for identification produces an over-built, expensive system. Designing every camera for monitoring produces a system that fails the moment something serious happens.
Lens, focal length, and field of view
A camera's field of view determines what's in the frame and how much pixel density falls on each part of it. A wide-angle camera covers more ground with less detail; a narrow lens covers a tighter area with more detail. Picking the wrong lens for the scene is the most common camera-system error.
Lighting
Lighting is the variable that determines whether the camera produces useful footage at the moment you need it. Backlight, glare, deep shadow, and very low light each have different solutions — and most can't be fixed after the camera is installed. Designing for lighting at the moment of greatest risk (after hours, parking-lot dusk, etc.) is what separates useful coverage from theater.
Mounting height and placement
Cameras mounted too high produce great wide shots and no usable faces. Cameras mounted too low get tampered with. Mounting decisions are a tradeoff between coverage, security, and the practical reach of maintenance. The right answer depends on the scene, not a universal rule.
Storage and retention
Storage decisions follow retention policy, not the other way around. Common retention windows are 30, 60, or 90 days, but the right answer depends on your incident review process, regulatory or contractual requirements, and your budget. Longer retention is not always better — fast, reliable access to recent footage usually matters more.
Network and bandwidth
Modern IP cameras can produce significant bandwidth, especially in higher resolutions. A camera system that fights with the rest of the network — or that gets bandwidth-starved during peak traffic — is a coverage problem disguised as a network problem.
Common mistakes
- Wide-angle cameras specified for identification scenes
- Outdoor cameras placed where they face directly into low sun
- Indoor cameras pointed at high-traffic doors with no lighting plan
- Camera count chosen before coverage goals are written down
- Retention specified without checking storage and bandwidth
- Cameras chosen for installation convenience rather than the scene
- No plan for who actually monitors or reviews footage
