My client is a 4 story, ground up, wood frame build of apartments. They want me to provide them with "fire watch" using video cameras. Traditionally fire watch is performed by a live guard on site in our area. I see several camera companies offering specialized smoke/fire detecting cameras on my Google search - any experience with them? Just wanting some opinions on how best to accomplish this. The site is rectangular and approx 350ft x 650ft perimeters. Thanks in advance for any input. PS: if anyone has info re: legal butt covering, I'd appreciate that as well. I'm a little worried anytime a customer asks me to watch for fire (we are not a life safety integrator)
Has Anyone Used Video To Perform Fire Watch?
I suspect the best approach is to talk with whomever is doing the firewatch monitoring (ie: remote video monitoring stations list) and ask them for guidelines and recommendations of which equipment to use.
Thank you. I did talk to my monitoring company and they recommended full time monitoring of an standard IP camera, preferably with PTZ so they can look around the entire site periodically. Way less expensive than an on site guard and it puts human eyes on all areas of the site all the time. They see no need for expensive special cameras as long as we let them watch them all the time. If we want to depend on alarming they were a little less positive and suggested high end technology and heavy CYA in the contract.
My guess is you should talk to FLIR. I wouldn't trust video analytics in the visible spectrum to work reliably.
I worked with a major integrator and saw our fire alarm division using these:
https://www.fike.com/products/fike-video-analytics/
Ugh...I know that stuff, cringe. On another note...here is a form of fire watch: http://www.alertwildfire.org/tahoe/firecams.html
Why are some of those aimed so high? Is it a privacy thing? I don't know the area personally. I get they want to watch the distance for fire, but I clicked through a few and the image was showing was half sky.
I saw several of these in use at museums when I worked for Smithsonian. I also saw the tests they used to verify that they were going to work, and they did. Of course, you still need a human or an auto-dialer to call 911 or the fire dept. to make it useful.
Who is requiring the fire watch? If this is being required by the local fire department or other AHJ or an insurance company, you better be sure that any solution is approved by them before proceeding. These folks are typically very conservative may have a hard time buying that a camera can replace a human in this application.
Exactly Michael. I would be amazed if the local AHJ would approve this unless this is alongside a traditional fire watch consisting of a guard or someone in person.
Apparently its the GC just concerned with their build. I will definitely make this cautionary note prominent multiple times in our proposal. We already use a disclaimer when talking about thermals detecting heat to make sure clients understand we're not saying they are a "fire alarm system". Thanks for the reminder.
I think that the automated smoke&fire detection may be considered only as an auxiliary tool for operators. The final decision must be accepted by an operator. Our software has been used (do not theat it as an advertisement) to detect forest fire. The customer installed PTZ cameras on top of cellular towers. PTZ cameras continuously patrol from a preset to a preset. At each preset the software tries to detect smoke&fire. If a couple of cameras detect smoke or fire at the same time we consider it as the same fire and perform triangulation to realize the real point on the map at which the fire takes place. The system works pretty reliable and AFAIK the customer greatly reduced the number of operators.
Hi Igor - how long do you have to leave the PTZ at each preset position to correctly identify that this is a fire?
How large does the fire have to be in a wide scene for the analytic to correctly identify it is a fire and not a guy lighting a cigarette?
how long do you have to leave the PTZ at each preset position to correctly identify that this is a fire?
As I remember the camera stays about 30 seconds at each preset.
How large does the fire have to be in a wide scene for the analytic to correctly identify it is a fire and not a guy lighting a cigarette?
Please refer to this example:
Thanks Igor
Concerning the image you sent : this seems like smoke detection. Were you detecting open fire or the smoke of the forest fire?
You are right. In this project our partner mostly detects smoke because it's too bad for a forestry if there is a visible flame already from this point of observation. But fire detection also runs on the same video channels. Size of flames should be approximately the same to be detected.
And... talk first with a security industry legal specialist, like K&K, Kirschenbaum.
Ask questions and get answers to your physical security questions from IPVM team members and fellow subscribers.