Video image fire detection, which “sees” fire or smoke rather than senses it, has become more popular due to its various advantages, including quicker detection time and ease of installation/maintenance. A thorough understanding of how the technology works and what components are needed is required in order to run a successful implementation.
Video image fire detection, which “sees” fire or smoke rather than senses it, has become more popular due to its various advantages, including quicker detection time and ease of installation/maintenance. A thorough understanding of how the technology works and what components are needed is required in order to run a successful implementation.
At the core of video image fire detection is the fire- or smoke-detecting algorithm, which can reside in the camera or on a server based on the system design. Basically, the analytics monitors movements of light patterns relative to a stable background. If the movement is consistent with known smoke or flame movement patterns, then an alarm will be issued. “Cameras trained on the area or process being monitored collect reference images, which are monitored by analytic software for changes of lighting across pixels that coincide with learned patterns that indicate the presence of dust, oil mist, flame, reflected flame, or smoke in the image,” said Rick Jeffress, Director of Business Development at Fike.
Video image fire detection needs a minimum of one camera and a system which is able to show the alarms. “Video-based fire detection is a customized solution. Depending on the requirements of the customer and the fire protection concept, additional components or connections can be included. For example, network video recorders based on iSCSI for alarm-based recording to ensure an easy root cause analysis,” said Sören Wittmann, Product Manager for Special Detectors at Bosch Security Systems.
In terms of the camera, it does not have to be high-definition as the analytics typically processes VGA resolutions. “Integrated algorithms in the camera make detection independent from the transmitted resolution without any influence on its fire detection capabilities,” Wittmann said. “For analytics, only the internal resolution of the camera is important. The higher the resolution, the better a camera is at covering larger areas.”
In general, the camera should be built to withstand certain extreme and tough environments. “There are off-the-shelf cameras that are ready for classified areas and extreme environments, such as the Moxa VPort-36 used with server based analytics,” Jeffress said.
Separately, it’s worthwhile to know that video image fire detection can act as a standalone system as well as a complement to an existing fire detecting system, verifying fire or smoke detected by traditional sensors. “In most cases where video is used as a smoke alarm verification application, the analytics will only run following a trigger from the fire alarm control panel,” said John De Blonde, Director of Product Line Management at Xtralis. “In the case of Xtralis SmokeTrace, once smoke has been identified, three image frames and one scrolling frame will be sent straight to the central monitoring station (CMS). The smoke will be highlighted by a bounding box, allowing the CMS operator to rapidly identify the fire. The CMS will then call fire and rescue services to confirm the smoke has been visually verified, triggering their response.”
Product Adopted:Other