What a megapixel is
A pixel is one point of recorded light. A megapixel is a million of them. A sensor with more megapixels records a finer grid, so the image can hold more detail and can be enlarged further before it breaks up.
Resolution is often written by name instead. 1080p is the common name for about 2MP, and 4K is about 8MP. Four times the pixels does not mean four times the detail in a usable sense, though. It means you can cover a wider area at the same sharpness, or the same area at a higher sharpness, depending on how the camera is aimed.
Pixels on target matter more than the MP number
Imagine two cameras looking at a driveway. One is a high-resolution camera with a wide lens that takes in the whole property. The other is a modest camera with a narrow lens that frames only the driveway. A license plate occupies a handful of pixels in the first and many more in the second. The second camera is the better plate reader, despite the lower megapixel figure.
That is the idea behind pixels on target. The sensor's pixels are spread across whatever the lens sees. A wider view means each object gets fewer pixels. A tighter view gives each object more.
Some specific targets have published guidance. For license plates, a commonly used rule is that the characters should be roughly 20 to 30 pixels tall for software to read them. We cover this in our LPR camera settings guide. Faces and clothing need their own minimums, and the right figure depends on whether you want to detect a person, recognize a known person or identify a stranger. Those goals need progressively more detail.
For how lens choice and view width change this, see our field of view guide.
Detect, recognize, identify
A useful way to think about resolution is by what you need the camera to do.
- Detect. Tell that a person or vehicle is present. This takes the fewest pixels and is where AI analytics work well on a wide view.
- Recognize. Tell whether it is someone or something you know, such as an employee or a company truck. This takes more detail.
- Identify. Pick out specifics that stand up in an investigation, such as facial features or a plate. This takes the most detail on the smallest area.
One camera rarely does all three across a whole site. That is why a layout of wide cameras for detection and a zoom camera for a close look is common.
The cost of more pixels
Higher resolution is not free. It shows up in three places.
- Bandwidth. A larger image makes a larger video stream. On a wired network that may not matter. On a cellular connection it can decide how many cameras you can stream at once.
- Storage. Bigger streams fill drives faster, or reduce the number of days you can keep. Our guide to how long security cameras should keep footage covers retention planning, and the cloud versus local storage comparison covers where the footage lives.
- Processing. Analytics, recorders and viewing software all work harder with bigger images.
Compression and frame rate also change the totals. Modern compression reduces size substantially, but heavy compression can smear fine detail, which defeats the point of paying for resolution. Higher frame rates add data but help with moving subjects.
Resolution and low light
More megapixels on a sensor of the same size means smaller pixels, and smaller pixels collect less light each. In dim conditions a very high-resolution camera can produce a noisier picture than a lower-resolution camera with a larger sensor. Resolution on a spec sheet says nothing about night performance, so judge a camera on sample footage from conditions like yours. Our night vision guide explains the lighting side.
Choosing resolution by task
| Task | What matters most | Typical approach |
|---|---|---|
| General awareness of an area | Coverage and reliable detection | Wide camera, moderate resolution |
| Detecting people and vehicles | Enough pixels for analytics to find the subject | Wide camera aimed at the area of concern |
| Recognizing a person | More pixels on the subject | Narrower view or higher resolution |
| Reading license plates | 20 to 30 pixels of character height, fast shutter | Dedicated, tightly framed camera on one lane |
| Close look at an event | High magnification on a small area | Optical zoom camera |
Resolution on a cellular connection
Sites without a wired network stream video over cellular, and that changes the math. Every camera on the unit shares the same connection, so a very high-resolution stream from every lens can crowd out the others or force heavy compression. A common approach is to send a lighter stream for live viewing and analytics, and keep full-resolution recordings for when someone needs to review an incident. Ask any provider how streams are handled when the connection is limited, and whether recording continues locally if the link drops.
Frame rate belongs in the same conversation. A busy lane of traffic needs more frames per second than a quiet storage yard, and each added frame costs bandwidth. Set resolution and frame rate together, based on what the camera has to catch.
Common mistakes
- Buying the biggest number. A 4K camera aimed at too wide a scene can still miss the detail you wanted.
- Counting on digital zoom. Zooming in on recorded footage only enlarges what was captured.
- Ignoring the network. Streams that do not fit the connection stutter or drop. Plan bandwidth before cameras.
- Judging by daytime samples. Check night footage and moving subjects.
How VDS approaches resolution
The VDS mobile surveillance unit splits the job the way this guide describes. Three 8MP wide-angle cameras cover the site, and a 5MP PTZ with 25x optical zoom closes in on detail, with optical zoom doing the work instead of digital cropping. Units connect over dual-SIM 4G/5G cellular, so no site internet is needed. If you are working out what resolution and coverage your site needs, talk to our team.