Efficient surveillance in the rail environment

Increasingly, CCTV systems are being used for ongoing surveillance, co-ordination of emergency services during an incident and post-event analysis by intelligence services. Furthermore, in the aftermath, much can be learnt from the recorded video footage. To be available during and following an incident, the critical parts of the system must survive the incident. Thus surveillance systems must be designed to meet a number of criteria.

System resilience

Designing a system that can survive a major incident involves examining all areas of vulnerability and where possible eliminating those areas considered a ‘single point of failure’.

The geographical layout of railways – plus the multi-level method of managing stations, lines and overall systems – means that the surveillance system should be designed to operate in a hierarchical manner. This hierarchical design facilitates the use of resilient systems and equipment, and therefore enables the system to be designed with incident management and recovery in mind.

Analogue-based surveillance systems use point-to-point cabling to transmit video, therefore a break in that path will stop the video transmission. Digital networks support many different paths for video and have the ability to re-route around a point of damage to maintain service.

The core IP network and all major sections should be designed to cope with average video throughput plus additional video from neighbouring sections, should a failure occur in that section. The network bandwidth can vary greatly between normal and emergency conditions, even if the network has not been affected by the incident.

Video requirements

Since a camera is considered to be a single point of failure, positioning is very important and should be chosen to provide overlapping coverage with adjacent cameras. Thus should a camera be rendered inoperable by an attack, adjacent cameras can be re-positioned to maintain full coverage of the area.

Traditional analogue CCTV provides high quality live video, but often offers very poor recorded video, making post-event analysis very difficult. Digital systems use networked video recorders (NVRs) where both recording and playback is of high quality. This allows much more detail to be viewed – for example, face recognition for suspect identification and tracking. It is therefore important to consider both the live viewing requirements for normal operation, and the recording requirements for post-event analysis.

The location of video recorders also has a significant impact on overall system design and its ability to survive an attack. The preferred solution is a partially distributed model, whereby the recorders are attached to the video network and located in secure areas, but are distributed around the network acting as collection points for each section. This means the network only has to handle the video for each local area, rather than all video travelling to a central point.

Intelligence and automation

The use of intelligence and automation in surveillance can significantly improve incident detection and prevention, and expedite post-event analysis and evidence gathering.

By combining intelligent motion detection with recording control, it is possible to only record from a camera when something of interest happens. For example, if a camera is positioned at an unmanned level crossing there is little point in recording the scene if nothing happens. However should a person or vehicle approach the crossing, motion is detected and recording starts. This combination of intelligence and automation substantially reduces the amount of data to be analysed following an incident.

Surveillance and video analysis can now be integrated with other systems such as access control, fire alarms and buildings management. This allows the system to monitor all aspects of the location and provide automated responses to predefined situations, as soon as a problem is detected.

Flexibility

A surveillance system should be both efficient and flexible. This is easily achieved with an IP network-based solution, which allows any operator to view any camera from anywhere on the network. It also allows remote monitoring if, for example, a local control room is damaged.

Remote monitoring allows those emergency services that are taking the lead to control the system in a hierarchical manner. Initially, fire and medical services may need control over certain areas to coordinate the rescue, then police and forensic teams may take over during the investigation and finally the repair and maintenance teams would come in.

After considering all these criteria, an IP network-based surveillance system capable of delivering continuous service in mission critical environments can be created. This surveillance system can also provide disaster recovery and incident management capabilities, facilitating the return to normal operating conditions as soon as possible.