A security camera system is specified around what you need to be able to prove, not around resolution. Accurate Security & Lock designs and installs camera systems for homes and businesses across Northern Illinois and Southern Wisconsin, matching lens, mounting height, lighting, recording and retention to each position.
Most camera systems fail at the same point: they record something, but not the thing you needed. A number plate that is unreadable at night, a face that is a smudge, or a camera aimed at a wall because it was quicker to mount it there.
We design and install security camera systems for homes and businesses across Northern Illinois and Southern Wisconsin, starting from what you need to be able to prove rather than from a specification sheet.
| What you need there | Rough pixel density needed | Lens and position | Typical mistake |
|---|---|---|---|
| Detect a person on a perimeter | 20 px across the subject | Wide lens, 2.8 mm class, camera can be higher and further back | Expecting this camera to identify anyone |
| Recognize a known person at a door | About 40 px | Medium lens, framing the approach path rather than the door slab | Covering the whole front garden in one frame |
| Identify a stranger at an entry | 80 to 100 px | Tighter lens, 4 mm class or longer, or two cameras on the same approach | Mounting high above the door so the face is foreshortened or in shadow |
| Read a plate at a driveway or gate | Depends on read distance and plate size | Purpose-built plate camera with tuned shutter and infrared, aimed level with the plate | Using a general-purpose camera and hoping |
| Watch a counter or till | Identification at close range | Close-mounted camera at face height over the transaction point | Mounting at ceiling height so hats and hoods hide the face |
| Cover a long boundary | Detection only | Several wide cameras, or analytics to filter motion | One camera rotated to cover everything, and therefore covering nothing |
Decide what you need to prove, then choose hardware
Detection tells you something happened: a person crossed the yard at 02:40. Identification gives you something usable, such as a face, a tattoo or a plate. They need very different amounts of detail in the image, and therefore different lenses, mounting positions and lighting.
Most disappointing systems were designed for detection and bought as if they would identify. The working measure is pixels on target.
- Detecting a human figure takes roughly 20 pixels across the body.
- Recognizing a person you have seen before takes around 40 pixels across the subject.
- Having a real chance of identifying a stranger, or reading a plate reliably, takes roughly 80 to 100 pixels. [Insert source: pixel-on-target thresholds for detection, recognition and identification]
Matching the camera to the position
Those numbers dictate how much of a scene one camera can cover and how far away it can be and still be useful. The choice follows from the position and what it has to deliver, not from a specification sheet.
Resolution is not the answer to bad light
A higher pixel count spread over a wider field of view does not add detail. What matters is pixel density on the subject, so a 4K camera framing a car park can put fewer pixels on a face than a 2-megapixel camera framing a doorway.
Resolution also cannot create light. A camera with a small sensor in a dim scene produces a soft, noisy image that looks fine as a thumbnail and is useless as evidence.
Lens choice and mounting height
A longer lens narrows the view and magnifies the subject, which is how you gain pixels on target without moving the camera.
Mounting height is a compromise. Higher is better for overview and less vulnerable to tampering; lower is better for faces. A camera at eight feet looking down a path sees a face; the same camera at fourteen feet sees the top of a head and the brim of a cap. Around doors, aim along the approach rather than down at the threshold.
Light, power and what happens after dark
Published infrared range is measured in ideal conditions and is optimistic in rain, fog or snow. More useful is sensor size: a physically larger sensor gathers more light, so it needs less infrared and produces less noise.
Shutter speed matters just as much. A slow shutter gathers light but smears anything moving, which is exactly what you are trying to capture. A camera claiming excellent night performance with a slow shutter and a small sensor gives you a well-lit blur where a face should be.
- An infrared image is monochrome, so it will not tell you what color a jacket or a car was.
- If color matters at night, add light at the scene rather than buying a more expensive camera.
- A switched or dusk-to-dawn light over an entry does more for image quality than upgrading the camera, and it deters as well.
Power over Ethernet and cable routes
Power over Ethernet runs power and data on one cable to a switch or recorder: one cable per camera, no separate power supplies, and battery-free operation. The practical limits are run length, around 100 meters or about 330 feet, and the switch power budget.
Cable routing is where installations are won or lost, and a route that cannot be installed tidily should be identified at survey stage rather than on the day.
- Cable routes that can actually be installed without wrecking a ceiling.
- Weatherproofing on exterior runs, with every penetration sealed, because exterior cable needs protection from water and UV.
- Camera positions that are not defeated by a passing van or a bush that grows back in June.
Retention, storage and who owns the footage
Retention is the number of days of footage you keep, and it is a straightforward calculation from camera count, resolution, frame rate and how much motion actually occurs. Thirty days is a common baseline, and some insurers ask for a minimum. The mistake is buying storage for the resolution you want and then discovering that recording continuously at that resolution gives you five days.
The system, the recordings and the credentials should be in your control, not the installer's cloud account. Set remote access up so it works from your phone, keep the recorder somewhere findable but not trivial to steal, and know who can view the footage and for how long. Footage that may be used in a dispute has to be exportable in a form that stands up.
When a lock or light is the better buy
A camera records; it does not prevent. At a side gate that is never latched, the cheaper fix is a self-latching lock. On a loading bay with a broken closer, the cheaper fix is the closer. On an unlit service door, the cheaper fix is light.
Where a camera is the only measure in place, it is being asked to solve a problem it cannot solve. We will point that out rather than quote for six cameras.
Service areas for security camera systems
We cover every town on our list across Northern Illinois and Southern Wisconsin. These are a few of them.
- Security Camera Systems in Fox River Grove, IL
- Security Camera Systems in Genoa City, WI
- Security Camera Systems in Grayslake, IL
- Security Camera Systems in Hebron, IL
- Security Camera Systems in Hoffman Estates, IL
- Security Camera Systems in Huntley, IL
Frequently asked questions
How many cameras do I need?
Fewer than most quotes suggest. The right number is determined by the entry points and the areas you need to identify people in, not by covering every part of the property. We survey first and propose coverage against your actual requirements.
Will cameras work at night?
They can, but night performance varies enormously. Infrared range, sensor size and the amount of ambient light all matter, and a camera rated for identification in daylight may only detect movement at night. We specify cameras for the lighting at each position.
How long are recordings kept?
Whatever you choose, subject to storage. Thirty days is a common baseline and some insurers require it. Retention is a straightforward calculation from camera count, resolution and frame rate, and we will size storage to match rather than leaving you to discover the limit later.
Is one high-resolution camera better than two lower-resolution ones?
Usually not. What matters is how many pixels land on the subject, not how many the sensor has in total. A single high-resolution camera covering a wide area spreads those pixels thinly and often puts fewer on a face than two cameras framing tighter views. Two cameras also give you a second angle, which is frequently what makes a clip usable.