The anatomy of a network camera
by Stephane Labrousse

Look inside a modern network video camera and you’ll find that it’s a far more sophisticated, capable device than a conventional analogue CCTV camera.

With its own embedded Operating System, IP server, signal processing software and even on-board memory it’s a true networking device, much like a real computer. As well as being able to respond to remote commands like ‘pan to these co-ordinates’, other camera functions can be controlled easily from a PC connected to the same network. This could allow quick adjustment of camera settings to optimise picture quality if lighting conditions change – all without the need to climb a ladder and make manual camera adjustments.

It can start recording automatically or send alarm signals when movement is detected in its field of view. It can transmit audio from an in-built or connected microphone through the same network cable as video and control signals. It can even function as an essentially autonomous device, recording images to its own on-board memory without occupying valuable network bandwidth.

Cameras like the Sony SNC-RZ30P combine superb picture quality with remote PTZ control, optional wireless network connectivity and advanced management functions plus an on-board memory option for autonomous picture recording in response to activity in the camera’s field of view.

Since it’s a true networked device, the camera can even benefit from standard IT industry wireless networking technologies. With a Wireless LAN (WLAN) card installed, a camera can be sited in difficult-to-reach, inaccessible or remote locations where it’s expensive or impractical to run additional cabling.

It’s true that these additional capabilities add to the unit price of a network camera compared with a conventional CCTV camera. But there’s a considerable payback to the end user when this additional cost is factored into the all-up cost of video network. Smarter and more capable than their analogue predecessors, network cameras shift a lot of activities to the camera – making IP-based video networks far more powerful, flexible, easy to expand and cost effective than analogue solutions.

In examining the capabilities of a network video camera, it’s also important to distinguish it from a consumer webcam. Aside from lacking the rugged build quality of professional network cameras, a webcam offers fairly limited picture resolution. Just as importantly, a webcam must be connected directly to a host PC to function: it’s essentially a single-user device, making it difficult to share pictures with other computers. A network camera, in contrast, does not require any additional hardware to operate and can be located anywhere a network connection is available, allowing images to be viewed by any authorised user on the network. Latest Sony cameras such like the SNC-P1 even feature an on-board ‘multi-cast’ function, simplifying the distribution of a single video stream to several connected PCs simultaneously.

Crucially, a professional network camera also has its own unique IP address – just like a server or other computer connected to the Internet. It’s this key distinction from analogue cameras that allows it to communicate with other devices on an IP network, including web browsers and network attached servers.

What can a network video camera do?
A network camera’s embedded IP server allows pictures to be monitored from any location – all that’s needed is a connected PC running a standard web browser. In other words, viewing pictures from a network camera is effectively the same as browsing a dedicated web site. Furthermore, the embedded server software within the camera includes its own ‘home page’ that can be used to set up and adjust camera functions via a connected PC. This approach means that there is an immediate cost benefit in choosing network cameras over traditional analogue designs. Since any number of cameras can be viewed and controlled via a standard PC, there’s no requirement for dedicated video monitors, camera multiplexers or other specialist hardware.

Viewing pictures from another connected PC is just the start. While a web browser such as Internet Explorer is all that’s required to see what’s in a particular camera’s field of view, installing dedicated management software – like Sony RealShot Manager – on the same PC turns it into a powerful camera management centre. A single user can view and control images from up to 32 connected cameras. It’s even possible to customise the display so that images from each camera can be overlaid on the schematic diagram of a building, making monitoring of a large, multi-camera installation far clearer and more intuitive for the operator.

As we’ve already seen, an IP server within the camera means that video pictures can be viewed from a PC that is connected to the same network. In the case of a large multi-national corporation operating its own Virtual Private Network (VPN) to link sites via the Internet, the possibilities become broader still. For instance, a networked PC could be used to browse and control images from a network camera in another building, at another site or indeed in another country anywhere in the world. Using standard IT industry encryption tools can guard images against electronic ‘eavesdropping’ or interception – making networked video monitoring a dependable option, even for organisations where security is mission critical.

More flexible camera management options
The embedded intelligence in a network camera means that it offers other benefits in terms of system management and control. Whereas CCTV systems are effectively under the command of a single observer/system manager – usually in a dedicated control room – video networks allow as many people as desired to view pictures via their desktop PC. This could mean streaming pictures from a theme park or other ‘web attraction’ on the Internet for anyone to see. Similarly, it allows every employee in an office building to view pictures from a monitoring camera at the entrance to a building on their desktop PC.

While anyone with a PC and web browser can theoretically view images from a camera on the same network, in practice there are sophisticated management functions in place to ensure that only authorised parties can ‘see’ picture. In practice this could mean limiting access rights to one privileged person – typically a security manager - or to everyone in a particular department. Sony cameras such as the SNC-RZ30P feature powerful security functions that include password protection and IP filtering to provide access to images and control functions for up to 50 simultaneous users connected to the same network.

Enhanced security applications
We’ve seen how IP-enabled network cameras simplify the configuration and management of video networks, but they also open up new exciting possibilities for enhancing the efficiency of security and surveillance applications.  

Many cameras – from high-end models to small, low-cost designs like the new Sony SNC-P1 – feature activity detection that can be triggered by an object appearing in the camera’s field of view. In response to this trigger, the camera will send an alarm signal via the same IP network… maybe alerting an operator to an anomalous event via an on-screen message or ordering a connected video server to start recording images.

Intelligent activity detection based on analysis of motion vectors enables the SNC-P1 to discriminate between genuine events in its field of view and false alarms caused by changing levels of illumination.

While this is an effective means of covering an area like a warehouse or underground car park where lighting levels are consistent, it can be harder to distinguish false alarms from real incidents in situations – such as out of doors – where light levels may be changing rapidly. A more advanced solution is offered by latest camera designs that feature motion vector detection. Powerful on-board signal processing enables the camera to analyse what’s actually happening in its field of view. This enables the camera to distinguish a significant event – such as a large object entering the picture – from changes in lighting across the whole scene caused by a cloud masking the sun.

As well as alerting the operator to an unusual incident taking place in its field of view, activity detection can also be used to trigger recording within the camera itself. Some cameras – such as the Sony SNC-RZ30P and SNC-Z20P – can accept an optional plug-in memory module capable of storing several minutes’ worth of images. This opens up the possibility of the camera serving as an effectively autonomous device. In the absence of activity in an area being monitoring, the camera will maintain a watchful eye without continuously transmitting a stream of video images, saving precious network bandwidth and disk space on a video server. In response to activity or an alarm event, however, the camera starts recording images to its on-board memory that can be subsequently examined or downloaded to a video server.

So far, we’ve discussed the role of the camera in capturing video images, but it’s also important to consider how operators can benefit from greater management and control possibilities once pictures have been recorded to a connected video server.

It’s tedious and time-consuming to search hours’ worth of material to find a particular incident, and indeed this has been a traditional limitation with tape-based recordings. In an IP world, however, the process of tracking down a critical incident becomes far easier. Network cameras have the capability to embed ‘metadata’ in the stream of video images sent over the network. This can typically include the time and date of a scene, plus other information such as the identity of the camera being used. It then becomes far easier to browse, search and review recordings by analysing this metadata – giving operators a far faster, more efficient alternative to examining hours worth of recordings in a library full of videotapes. In the future, these metadata capabilities will become more powerful still, allowing cameras to transmit scene-specific information such as the speed and size of moving objects in their field of view. This opens up the possibility of exciting new ‘smart’ security applications – giving security professionals a far more detailed, informative picture about what’s really happening than analogue CCTV systems ever can.