The camera is a visual spectrum detector, it does not use heat, UV, or IR to detect smoke or flames. The camera has a Texas Instruments Digital Signal Processing (DSP) multi media processor on board the camera, all the detection algorithms are embedded on this chip. The processor scans and reads every pixel on the image looking at the luma (or brightness) value that ranges from 0-256, there are 640 X 480 pixels per frame, or over 300,000 data points that are analyzed at 15 times per second for over 4.5 million data points per second. The flame algorithm DSP filter is looking for a group of core bright pixels with a group of pixels around the core flickering at the frequency of fire. Once a pattern of interest is identified it is fed through a feed forward neural network that has been trained on literally thousands of fire patterns and nuisance events to determine if the pattern is indeed a fire. Once the algorithm has triggered for 4 consecutive seconds it sends video alarms to the remote monitoring system and activates a dry contact.
There are 5 detection algorithms on the camera. Two algorithm detects flame, one detects flame directly and one detects through reflective firelight (the flame is blocked but reflects off surfaces). Two algorithms detect smoke and one detects motion. All algorithms work independently and can detect multiple events simultaneously within the field of view. The camera also contains algorithms that can identify trouble and supervisory conditions such as loss of picture, poor image, no light, a system crash and/or loss of power.
This video is the SigniFire IP network camera detecting a large pan fire at the UL large fire test facility in Chicago. The red box indicates the camera has identified the fire. The fire was set as a demonstration fire for insurance and end users to demonstrate the extinguishing capabilities of water vs AFFF, as well as the SigniFire system.
This video is a small 6 in pan fire with rubbing alcohol. The fire took place in a air Hangar in MD, the owner was an avid hunter and wanted to protect the animals and plane in the facility. Again the red box indicates a detection by the SigniFire Camera.
This is a video of reflected fire light being detected. The fire source is a 1 sq ft pan fire behind construction material in our warehouse. The camera is 60 ft. away. The purple tiles represent the time of alarm.
This video is the SigniFire IP camera detecting a Regin safe smoke emitter. The blue line indicates that smoke has been detected and this would be communicated to a fire alarm panel and/or the SpyderGuard monitoring station. The detection occurs much sooner than any type of conventional detector (spot, beam or air aspiration). The demonstration was done at a polyurethane manufacturing facility. The room is rather large with ~50 ft ceilings and is very hot due to the process so smoke stratification is an issue as seen in the video.
This video illustrates our motion detection algorithm. As the truck passes by you will notice green boxes indicating motion and producing an alarm event. The video image can be zoned and scheduled for motion detection. For example a motion zone can be placed around a door and attached to a schedule so anyone using the door between 7 pm and 6:30 am would be recorded, or initiate and alarm.
These are the IP network camera specifications, the detection capabilities listed are those that were tested as part of FM approval and our UL listing. The camera has demonstrated the ability to detect sources at greater distances (flame at 300 ft and smoke at 350 feet) but the fire size and amount of smoke has to increase proportional to the distance away from the camera. The camera uses standard security components such as the TI chip that is used in many high end security cameras and HDTV’s
SpyderGuard is a powerful tool used for alarm management, event re-creation and installation and commissioning of the system. Live images, floor plans and archived events can be viewed. The cameras setting and properties windows can also be accessed for installation and commissioning.
SigniFire system architecture.
Once a fire event is detected, there must be a written and rehearsed response plan
Image of Pelco explosion proof housing. The SigniFire IP camera can detect threw glass and will fit in almost any CCTV enclosure. We have used stainless steel for marine applications, indoor and outdoor housings, dust proof and water tight enclosures. The next video is of a unauthorized welding event with the camera in the above class 1 div2 rated Pelco enclosure.
Video of a unauthorized welding event in a coal tripper room at a PRB coal plant. If the welding was authorized a masking zone could be placed around the area in the image where the welding is to occur and attached to a schedule. The SigniFire camera could then detect events outside the masked area.
SigniFire is targeting large volume structures with high valued assets where there are no real detection methods that are effective in these type of spaces.
This video shows smoke detection at a warehouse/distribution center as it drifts down an aisle
This video shows SigniFire detecting a 1 ft fire at Dominion Energy plant in Kincaid Illinois.
SigniFire detecting smoke at Dominion Energy. Smoke started below the turbine deck. The situational awareness that SigniFire offers threw the live video is of great value in these industrial and manufacturing facilities because the operators are able to see what is occurring in the space and take the appropriate actions without sending someone into the plant and a potentially dangerous situation to see what is going on.
Outdoor detection of fire, Infra red cut filter from camera is typically removed for outdoor applications
Representative SigniFire Customers
We recently installed a system in a Hanger for Sikorsky with plans to install additional systems in transmission test bays at the CT facility.
This is a video of a smoke detection at a waste to energy plant in CT. It is a very difficult environment due to the large amount of dust and debris in the air and steam that can come off the garbage from the decaying process.
SigniFire detecting hot smoke at University of Maryland’s Cole Field House. The camera is about 300 ft away from the smoke. In a structure like this, there is no effective detection available.
Nuisance alarms can be mitigated during the commissioning process which takes about 1 to 4 weeks. The process can be done remotely. Initially all events are recorded and reviewed to determine the best way of eliminating them. This can be done through a combination of adjusting sensitivity levels, creating zones to block areas of the image, and/or incorporating delays into the algorithms and/or dry contacts.
The top image on the left illustrates an alarm caused by steam release on a boiler. The steam release was expected and Marriott did not want to detect the steam in the future. So on the bottom image a smoke block zone was created (blue rectangle) so the system would not alarm in the future to a steam release unless it breaks out of the zone.
We recently conducted a comparative test series with Schirmer engineering. We tested ion and photo spot detectors, beam detector, linear heat, and air aspiration system along side a SigniFire system in a high bay environment (18ft).
4 Standard UL sources were used during testing (UL wood crib, UL Smoldering wood, UL 6 in pan fire, and UL tamped paper test). We also included a smoldering cable bundle, electrical fire are the number one cause in manufacturing facilities, and two smoke emitters (black and white). The smoke emitters are very repeatable sources and we use them during system commissioning.
Above is a summary table of results. The Schirmer report foes into a lot more detail and we can supply the report at no cost. The conclusion determined from the testing is that the SigniFire VID system is just as quick if not quicker than any other detection method for these small fire sources in areas where the ceiling is 18ft or greater. The only other system to detect 100% of the fire was the air aspiration system with the sensitivity increased to 0.2%/m and even with the high sensitivity was still outperformed by SigniFire 88% of the time.
Above is the characteristics that we look for. End users and consultants should consider VID systems when they have a large volume space, desire for security and or a high values asset. In addition the fire source has to be visible to the human eye. SigniFire cameras will not detect a few fire sources such as pure hydrogen flames that produce no visible flame. Because of the fast detection capabilities of the SigniFire cameras Trained onsite personnel are a plus. They provide the ability to suppress the fires at the time of detection when the fire is small and can be extinguished by a hose line, fire extinguisher, or by something as simple as removing the power source to an over heating appliance.
Examples of some SigniFire system designs. The designer should consider the vertical and horizontal orientation of the camera and the resulting field of view when placing cameras to cover the hazard area. Basic CAD files are available to provide a stating point for designs.
SigniFire at this time is addressing the non-residential structures market. In the latest statistics published by United States Fire Protection there was $2.8 billion in commercial property loss in 2000. The leading cause of fire was arson and 72% of these structures had no smoke alarms present. These facilities typically have high ceilings and may be sprinklered but do not install fire detection devices.
Fire Protection is like an insurance policy, it really does nothing for you and is an expense unless you have an event. In order to justify additional early warning protection you must perform a risk analysis. Annual risk expense of fire must be calculated. If SigniFire can mitigate this expense then it could be a justified capital expense.
Example of a basic risk analysis for a warehouse
An excerpt from an article written by an insurance consultant for Society of Fire Protection Engineering Magazine
The above outlines the advantages of SigniFire IP network cameras.
SigniFire Video Image Smoke
and Flame Detection
Fike Video Image Detection
What is Video Smoke and Fire Detection?
• Uses visible spectrum
spatially resolved video
images from stationary
• Analyze each frame pixel
by pixel using Digital
Signal Processing (DSP)
high and low pass filters
• Converts 15 frames of
pixel data to one
• Pattern is fed to neural
network for confirmation
• Video Analytics
– Artificial Intelligence
– Neural Networks
– Reflected Fire Light (RFL or Offsite)
– Motion Detection
– Image change
– Loss of camera
– Poor image (dirty, low light, out of focus)
– System Failure
– Loss of power
• Remote monitoring software application
• Functions on a standard PC
• Provides video surveillance, fire safety
monitoring, and allows integration of an
unlimited number of SigniFire systems
• View active alarm conditions
• Receive an audible voice warning
• Locate the event on building and site
• Coordinate response efforts
• Continuous 24/7 Video Recording
• Alarm events archived for post event
• SigniFire IP Network cameras
• CAT 5 Computer Cable
• IP Infrastructure (Switches)
• FSM-IP Network Video
• Remote Monitoring Software
• Dry contacts
• Monitor Module
• Fire Alarm Panel
• Call Fire Department
• Start Notification
– Pull station
• Attempt Suppression
• Smoke Control
• Record Response
• Report to
• System Modification
mitigate future false
Company Policy – Written procedures – Training Plan
• Conventional detection methods are not suitable for
many of the environments requiring protection
• The nature of these detectors makes the severity of a
nuisance alarm greater due to limited situational
• Codes mandate that fire equipment be installed for life
safety and to prevent conflagration
• DOE Hazard Analysis
– FHAs for high bay locations should consider the effects of
smoke/hot gas stratification that may occur at some intermediate
point below the roof or ceiling as well as the potential for delayed
Smoke Detection in a Distribution Center
Comparative Testing Objective
• Compare the smoke and fire detection
capabilities of various technologies during the
developing stage of a fire in a high bay
• Testing was witnessed and documented by
• Equipment was designed and installed by fire
• Test Facility measures ~60 ft by 35 ft with 18 ft
Detection Technology % activation (% first to Activate)
ASD Level 1 100% (6%)
ASD Level 2 70%
Ion 55% (6%)
VID 100% (88%)
Development of Design Fire Scenario
Facility Characteristic –
• Large Volume Space
• High Value Assets
• Desire for Security
• Outdoor flame detection
Fire Characteristic -
• Visible Signature
Occupant Characteristics –
• On Site Security /
Risk Management of Fire
The Risk of Fire is equal to
Probability of a Fire
Loss associated with that Fire
Warehouse Fires - Example
Annual Risk Expense –
Probability of Fire X Single Loss Occurrence
• 1% chance of fire that requires fire brigade intervention Avg. cost $590,000
• 0.25% chance of warehouse fire that causes roof collapse Avg. cost $6,000,000*
• Average annual risk expense = (.01 X 590,000 + .0025 X 6,000,000) = $20,900
*The Probability of Fires in Warehouses and Storage Premises, Hymes & Flynn, 1992
SigniFire mitigates this risk through very
early detection and situational awareness
4 SigniFire Camera System can be
installed for less than $20,000!
XL Gaps Insurance article excerpt from
The $250,000 figure is a representative loss tolerance established by the
insurance company. In many cases, properly designed sprinkler systems
will satisfy this objective. Loss experience is usually the guide as to whether
or not a properly designed sprinkler system will meet this objective. If the
occupancy is particularly smoke sensitive, has a high business
interruption potential or has an extremely high value-per-unit volume,
additional protection may be needed.
At this time, specific performance criteria to prevent excessive smoke
damage are usually lacking. There is no agreed upon criterion such as
reducing smoke concentration to some specified value. Instead, more
qualitative solutions are used such as ignition resistance of wet benches,
early smoke detection, faster suppression of smaller fires, smoke
control or separation. If these measures are not felt to be adequate, an
underwriting solution such a proportional share policy may be used.
• Combustible dust in a facility can cause a very
dangerous condition in suspension
• SigniFire can detect dust clouds that resemble
• Several of the largest explosion and fire events
were attributable to dust
– West Pharmaceutical
– Imperial Sugar
– Grain Elevators
– Coal Dust
Maintenance and Commissioning
• Report generation
• NFPA requirements
– Left to Manufacturers
• How to commission
– Commissioning report
– Ways to test system
• Propane torch
• Smoke emitters
• Canned smoke
• Motion zones
• User initiated alarms
• Volume Detection
• Large coverage area
• Situational awareness
• Continuous video
• Standards and Listing
• Fast detection
• Nuisance rejection
• Post event
• NFPA 72, FM and UL
Provides a solution to fire and smoke
detection in large volume facilities where
no detection has historically been deployed
• Questions, Comments, Words of Wisdom
• Contact Information
Fike Video Image Detection
47 Loveton Circle, Suite F
Sparks, MD 21152