A 'smarter' approach to smoke detection

July 19, 2019
Smoke Detective leverages cameras in cellphones, smart devices to provide a visual smoke detection solution

One of the most fundamental technologies historically deployed in both commercial and residential security installations is smoke detectors and alarms. After all, the chances of dying in a fire where smoke alarms are present are significantly less than when they are absent as evidenced by the statistics.

According to research published by the National Fire Protection Association (NFPA) earlier this year, almost three out five home fire deaths from 2012-2016 were caused by fires in properties with no smoke alarms (40%) or smoke alarms that failed to operate (17%). In addition, the NFPA found that the risk of dying in reported home structure fires is 54% lower in homes with working smoke alarms than in homes with no alarms or none that worked.

While no one would argue that smoke alarms and detectors save lives, the technology itself – prone to failure without proper maintenance – has remained relatively unchanged for years. One company looking to change this is Smoke Detective, which has developed software capable of turning the cameras in smart devices into smoke detectors.

Steve Davis, the company’s founder and CEO, says he first got the idea for the product after lightning struck his home during the early morning hours one day in 2005. Although they had working smoke alarms and everyone was able to make it out safely, Davis says the incident made him more fire safety aware and after doing some research, he was horrified to learn just how many smoke alarms fail when they are actually tested.

Once it became apparent that smartphones would forever alter the communications landscape, however; Davis knew that extending the capabilities of traditional smoke alarms to something that could be embedded within one of these devices was the route he wanted to take. “As smartphones got more and more capacity, I though well maybe I could just create an app, for lack of a better word, for a smartphone so that your smoke detector could always be with you,” he says.

Davis subsequently teamed up with Dr. Gustavo Rohde, who was running a digital imaging lab at Carnegie Mellon University at the time, and they soon began work on what would become Smoke Detective.

How it Works

Unlike ionization, photoelectric or combination sensor smoke alarms that are commonly deployed throughout most homes and businesses, Smoke Detective leverages the cameras embedded in today’s cellphones and smart devices and continually analyses those images to alert users to the presence of fire or smoke within a room. 

According to Rohde, who current serves as an associate professor at the University of Virginia and as Chief Technology Officer for Smoke Detective, what differentiates the company’s technology from other video-based smoke detection solutions on the market is the installation, calibration and operator training process that is required to make these products work which is on a much higher level and scale than what is required for Smoke Detective.

“You’re probably talking about something like an oil refinery, power plant, a large warehouse or something specialized that would require or justify this kind of more expensive and extensive kind of installation process,” Rohde explains. “Some of these solutions also require more advanced hardware with infrared imaging capabilities, for example. What we’re not knowledgeable of is a solution that fits in a cellphone that anybody can use and doesn’t require training, calibration, or any installation in the sense that it has to be mounted or connected to some other device or a process where you can just go to the App Store or Google Play Store, download something and you have a working smoke detector. That doesn’t quite exist as far as we are aware of.”

For the solution they hope to bring to market within the next few months after they’ve completed testing, Rohde says they are intentionally using the lowest quality image data to ensure that it works with most of the cellphones that are prevalent in the market.

“For example, we’re testing with a lot of different models of iPhones right now and they have different kinds of cameras in them,” he says. “The programming interface that we have to connect to these cameras to get data from it allows it to be of different quality and we’re using the lowest quality right now to make it work and the reason for this is two-fold: one is, primarily, we want to have the lowest system requirements so that we can take this algorithm we have and just put it pretty much anywhere. Second, we also want it to be a light algorithm in terms of computation. We don’t want it to require large computation because the more pixels you have the more computation you have to do. We think we can make it work with a very low resolution image.”   

When it comes to detecting smoke in the dark, Smoke Detective is aided by IR illuminators where available but in cases where that is not an option, especially with many cellphones, Rohde says they can also take advantage of LED lights which are found with most cellphones to provide flash illumination for photos. Most of the latest generation smartphones, however, particularly those that can be unlocked via facial recognition, are IR ready.

Moving Beyond Smartphones

Davis has been working with Rohde to bring Smoke Detective to the market for the last five years and the company recently received a patent for its software in May. The company, which recently received funding from Innovation Works to further develop the technology, was also recognized earlier this year as a CES 2019 Innovation Awards Honoree.

While the initial proof concept for Smoke Detective was successfully performed on a smartphone as it was an easy platform to develop for, Davis says that it can be utilized with a wide variety of other devices with embedded cameras, including baby monitors, computers and security cameras. In fact, Davis says they are currently in talks with a “large security company” that is interested in licensing Smoke Detective’s technology and working with them to develop a variant of the algorithm for their cameras.       

“We’ve proved the concept so now we need to develop it for specific platforms and that’s what we’re doing,” Davis adds. “We think the low hanging fruit would be security companies who are doing verified video monitoring and those types of applications where it doesn’t necessarily have to be a certified life safety device but could be backup verification. Ultimately, we will be seeking UL certification, etc. but right now we've focused on spending our resources developing the technology.”

Ideally, Davis says he could envision end-users having all three smoke detecting technologies – photoelectric, ionization and visual analysis – installed within their homes and businesses and that the visual component could take many different forms.

“If you think about your house, how many different devices do you have that have a computer and a chip – your smart TV, laptop, desktop –and that’s down the road but the beauty of software is that the incremental cost, once you develop it for a platform, is not a lot so it could become a platform technology you see in a lot of different places making the world safer,” he says.

About the Author:

Joel Griffin is the Editor of SecurityInfoWatch.com and a veteran security journalist. You can reach him at [email protected].