10.12.2018 Views

RiskXtraDecember2018

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

x<br />

RISKXtra<br />

CCTV Cameras and Fire Detection:<br />

Moving Towards Integrated Solutions<br />

Video fire/smoke<br />

detection is a safety<br />

solution based on the<br />

computer analysis of<br />

images provided by<br />

standard CCTV<br />

cameras linked to a<br />

self-contained<br />

processing system<br />

capable of recognising<br />

smoke and flame<br />

patterns within the<br />

video image. Here,<br />

Risk Xtra interviews<br />

the Fire Industry<br />

Association’s CEO Ian<br />

Moore on the subject,<br />

duly focusing on its<br />

invention, subsequent<br />

development through<br />

time and the future of<br />

detection using<br />

surveillance cameras<br />

Ian Moore: CEO of the Fire<br />

Industry Association<br />

Risk Xtra: Looking back to when you first<br />

brought video smoke detection solutions to<br />

market all those years ago, what was the fire<br />

safety situation at that time? What was the<br />

initial ‘driver’ that really underpinned the<br />

desire to converge security equipment with<br />

fire safety products?<br />

Ian Moore: The mid-to-late 1990s was an era<br />

that witnessed the explosion of digital signal<br />

processing, and specifically towards image<br />

processing. There were numerous algorithms<br />

around to detect motion within the scene of<br />

view of a given CCTV camera which rapidly<br />

became a de facto offering as an intrusion ‘bolton’<br />

through DVRs and/or smart cameras, etc.<br />

By today’s standards the method employed<br />

was quite crude, using as it did a percentage<br />

contrast change pixel by pixel. There were even<br />

a few basic detection algorithms looking at the<br />

motion patterns of smoke, but none that were<br />

much more than conceptual or laboratory-run<br />

research projects.<br />

The only fire safety situation back then that<br />

allowed new technology to be used was the<br />

acceptance of engineered solutions for areas<br />

that couldn’t be suitably covered by<br />

prescriptive traditional means such as smoke<br />

detectors and beam detectors, etc. For their<br />

part, designers would rarely ‘stick their neck<br />

out’ on new and uncertified products.<br />

In the world of integrated fire and security<br />

systems, this was the first product that sat<br />

firmly in the middle, using the equipment<br />

backbone of a CCTV security system for the<br />

purposes of fire detection.<br />

Risk Xtra: Winding the clock forward to the<br />

present, what are the newest developments<br />

in the realm of video fire detection?<br />

Ian Moore: Technological advances in computer<br />

processing power and memory have given us<br />

more opportunities to increase the performance<br />

and uses of pretty much the same technology.<br />

Faster processing allows us to make many more<br />

calculations with the same data, while higher<br />

frame rate analysis and enhanced resolutions<br />

of camera images give us more accurate data.<br />

IP networks and smart cameras (ie on-board<br />

processing) allowing decisions to be made at<br />

the camera – which is certainly different from<br />

the initial days of using video ‘grabber’ boards<br />

in a PC and a DOS operating platform – boosts<br />

system design options.<br />

There’s often more than smoke emitted from<br />

a fire (or a potential fire). Systems that<br />

implement a range of detection for pinpointing<br />

smoke, flames, visible heat (through infrared),<br />

audible and gases (through spectral analysis),<br />

etc will yield a wider range of detection as well<br />

as the potential to detect fires earlier and thus<br />

reduce false alarms.<br />

A self-learning algorithm could assist in<br />

building up a history of what to expect at<br />

certain times of the day (in terms of lighting<br />

angles/intensity, how busy is the scene and<br />

different processes, etc). That said, how this fits<br />

into the standards landscape right now is<br />

difficult to see.<br />

There are a number of quality companies<br />

around the world working on this technology.<br />

Some of them have been investing heavily in its<br />

development for nearly 15 years. I hope that<br />

they continue to have the vision – if you’ll<br />

pardon the pun – of what video detection can<br />

be in the future once the performance<br />

standards have been agreed and video can then<br />

be an accepted mainstream form of detection.<br />

Risk Xtra: What are the main challenges to<br />

be faced when integrating video-based fire<br />

detection into a security, building<br />

management or alarm system?<br />

Ian Moore: Much of it is about mindset (ie ‘fire<br />

is fire’ and ‘security is security’). Rightly, there<br />

are fears around the corruption of signals, and<br />

especially so over IP networks. Although these<br />

can be secured, the reticence remains.<br />

Another big reason is product/system<br />

approvals which, in many cases, don’t allow<br />

usage in combined systems. Even though the<br />

technology has been around for over 20 years<br />

now, it’s still considered to be in its infancy.<br />

Risk Xtra: What are the benefits of videobased<br />

fire detection in comparison with<br />

regular fire detectors?<br />

Ian Moore: Without wanting to give a sales<br />

pitch, there are plain facts to state. For one<br />

thing, there’s faster detection as the camera is<br />

usually looking at the source of the fire and not<br />

waiting for emissions to reach a detector. Also,<br />

you can use an existing CCTV system. As such,<br />

costs can be reduced (plus there’ll be limited or<br />

zero interruption to the site during installation).<br />

Visual verification of alarms is another core<br />

benefit. With an ever-increasing requirement for<br />

56<br />

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!