IP Cameras

‘IP’ stands for Internet Protocol, so an IP camera is an Internet Protocol camera, which means that the camera has an IP address, which is unique for each device, and can send and receive data over computer networks and the Internet. Therefore, IP cameras are also often referred to as network cameras, or IP-based or network-based cameras. These terms are used interchangeably.

For a simple explanation, you can also describe an IP camera as a ‘computer with an image sensor’. That is, an IP camera, like a computer, has its own CPU, internal memory and (often) on-board storage capabilities. The image sensor captures image data and this data can be processed or analysed by on-camera software before video data is sent to a location elsewhere on the network or the Internet.

With computer chips becoming ever smaller and more powerful, IP cameras are benefitting from an increasing amount of CPU capacity to run more powerful algorithms that can undertake increasingly complex processing and analysis of the image data they receives from their image sensors.

As a real-world benefit, this means an IP camera can enhance the quality of the video data it receives before it is sent out, plus it can detect objects and patterns in video images that enable the camera to send alerts or trigger actions when pre-defined situations are happening. This then is where we get into video analytics.

Yes, a camera with a built-in infrared (IR) LED is able to capture images in complete darkness. However, the images captured during darkness will be in mono colour (black and white) only.

Today’s advanced technologies have resulted in more and more security cameras becoming increasingly capable of capturing colour images in what’s often described as ‘near darkness.’ Cameras are increasingly outperforming the human eye when it comes to seeing in the dark.

Light sensitivity in CCTV cameras is stated in LUX. In theory, the lower the LUX measure the more light sensitive the camera is. This single measure is useful in a small, controlled environment but not so much when a camera is covering a large outdoor space with scene complexity and environmental conditions playing a part. Therefore, in many cases, you cannot just simply rely on the LUX information in a data sheet.

No, the right resolution depends on a number of factors, including:

1. What is the purpose of the camera, what are the operational requirements?
2. What will be the maximum distance from the camera to the 'region(s) of interest' within the camera view? What is the size of the area you are looking to cover?
3. What are the light levels available and is any additional illumination required?
4. Where will the camera be positioned, and how much will it be tilted or angled?
5. How much network bandwidth and storage capacity do you have available or do wish to allocate?

Naturally, customers tend to play it safe and are drawn to choose camera resolutions that may be unnecessarily high for their use case, potentially spending too much on camera equipment and/or wasting bandwidth and storage resources for the life of the camera. The latter can be particularly wasteful and costly in large camera systems.

In practice, and as a rough rule of thumb, the use of an IP camera with a resolution of HD 1080p / 2 megapixels, and featuring high light sensitivity and enhanced WDR, will satisfy the operational requirements in the majority of cases. It is when long-range monitoring or minute detail is required that higher camera resolutions will be needed.

If you want to get it right, look at 'pixel density' instead of resolution. Pixel density is the number of pixels per centimeter (or inch) captured by the camera from the objects, persons, or vehicles in the scene. You can then use a table, such as the one below for monitoring people, to decide the number of pixels required to meet your operational requirement, and choose the right camera.

We have a blog post that shows some good examples of what different pixel densities will look like in real, go to Useful Examples of Pixel Density in Network Cameras.

Frame rate is expressed in ‘frames per second’ (FPS), or sometimes ‘images per second’ (IPS). Both are the same thing.

Most IP cameras offer 25fps or 30fps, but there are also high-speed IP cameras providing 60fps, 120fps, or even 180fps for exceptional use cases. But the frame rate capability of cameras increases as technology continues to evolve. Note that higher frame rates often can only be produced at a reduced resolution setting.

When deciding on frame rate, the general, simple rule is that fast activity requires a higher frame rate, and a slower frame rate is fine for slow activity. For example, a camera used for monitoring a fast moving production line in a manufacturing plant may be set to 60fps or more, whereas a camera monitoring a business reception area could be set as low as 5fps.

The correct frame rate is very use case specific.

For CCTV cameras that are monitoring areas where people are walking or, for example, vehicles are entering a car park and drive slowly, a frame rate of between 5fps and 10fps is often suitable.

Don’t forget that you will be wasting network bandwidth and storage resources if you set a high frame rate unnecessarily. This can be costly in the case of large CCTV systems.

The frame rate comparison video below will help you visualise what different frame rates look like.

What is the correct frame rate for your CCTV system?

The advantage of fixed lenses is that they are (1) less costly, and (2) smaller. Therefore, you will see that IP cameras with fixed lenses generally will be cheaper and more compact than cameras with varifocal lenses.

However, since a fixed lens cannot be adjusted, whatever the camera captures is what you will be given. Though, note you often can manipulate the image and crop or cut out that part of the image you’re not interested in.

Lenses, particularly good ones and varifocal types, are expensive and form a significant part of the cost of a camera. This explains the significant price differential between CCTV cameras with a fixed lens versus cameras with a varifocal lens.

Varifocal lenses have parts that can move in and out, and therefore require more space within the camera enclosure, hence varifocal cameras are almost always larger than fixed cameras.

The big benefit of varifocal lenses is that you can precisely set your camera view and capture the area you are interested in (much) better detail. This is particularly important for monitoring people, vehicles or activity at a greater distance, or when the application requires clear detail, which is often the case when using video analytics, such as face detection or ANPR installations.

In simple terms, video analytics is the capability of a camera or a computer to analyse the data captured by a camera and to recognise objects, activity, or patterns. The algorithms that make up the video analytics capability are ‘trained’ using large amounts of data sets that give them this ability to ‘recognise’.

For video analytics to work well, it will require the algorithm to be trained well, and for sufficient computing power (CPU) to be present for the camera or computer to be able to run the algorithm. These two critical prerequisites will determine how useful and accurate video analytics in a camera system will be in real-life situations.

It is, therefore, important to recognise that the usefulness of video analytics can vary greatly between different manufacturers. Generic phrases such as ‘video analytics’, ‘intelligent video’, and ‘AI’, etc., are widely used by marketers in the CCTV industry, but, as with everything, performance levels may differ significantly. Being cautious and taking expert advice is recommended.

The key benefits of video analytics are fourfold:

1. Accurate detections, and a significant reduction of false alarms
2. Undesirable situations or activity can be highlighted to operators or managers in real-time
3. Video evidence can be found swiftly during post-incident investigation
4. Processes can be automated, e.g., an ANPR camera can open a barrier if the number plate is recognised

For video analytics to function well, camera positioning, pixel density, and light conditions are all key considerations to be made as part of your planning.

Contact us if you require assistance with deciding what’s the best solution for you.

IP cameras have on-board CPU, internal memory, and often external storage capabilities (via an SD card), and could be seen as a computer with a sensor attached for capturing images. IP cameras can execute files, and a hacker getting access to one or more IP cameras on your network could create a powerful launch platform for a serious attack onto your organisation.

In addition, for someone being able to access a camera and/or its recordings, will instantly pose a threat to your organisation, create a security breach, and also violate data protection laws and regulations.

Deploying cybersecurity best practice needs to be a careful, serious consideration when using IP cameras.

Contact us if you require assistance with cybersecurity best practice.