Fire is complicated and therefore fire protection is complicated. There are no silver bullets for protecting people and assets from fire, so fire protection solutions employ various prevention, detection and suppression methods in combination to minimise the risk of fire disaster. An estimated US$33 billion was spent on fire protection systems in 2013, which is expected to grow to US$80 billion by 2020.
Fire protection systems are comprised of a suite of fire prevention, detection and suppression products and services that are selected based on the nature of what is being protected, be that factories, homes, shops, laboratories, mines, farms or forests. Furthermore, fire protection technologies are constantly evolving, offering increasingly effective ways to minimise fire disaster. Those responsible for fire protection seek out better solutions while balancing budget constraints on a continual basis.
Wildfire protection is no exception. Protecting forests from wildfire is complicated by terrain, fuel types (i.e. types of trees, plants and dead plant material that is prone to burning), accessibility, proximity to human activity and innumerable other factors. Wildfire prevention solutions range from public awareness campaigns to fuel management to fire modeling and prediction. Wildfire suppression can involve firefighters, airborne firefighters, fire trucks, helicopters, water, dry chemicals, building fire breaks, not to mention all the hoses and equipment that is involved with these. Even wildfire detection can require a relatively complicated solution.
There are two primary mechanisms being used by forestry personnel to detect wildfire: smoke detection and heat (or thermal) detection. Smoke detection is the most conventional way to detect fires. Humans in watchtowers search for smoke to indicate where there might be a wildfire. The semi-automated smoke detection systems available today use CCTV in combination with smoke-identifying algorithms to extend the viewing range of a human to be able to detect fires. Thermal detection uses modern thermal imaging sensors to identify the heat from a fire. Heat is created much more quickly than smoke, so fires can usually be spotted and suppressed more quickly. However, this modern technology does have its limitations. Here we give a comparison between smoke detection and thermal detection, analysing their respective strengths and weaknesses.
The key advantage of smoke detection over thermal detection is that it doesn’t require line-of-sight to the fire. In other words, if a fire is behind a mountain, it can be spotted by the smoke column it will eventually generate. Another advantage of smoke detection is that, in the case of humans in a watch tower, the range of detection can be up to 30km. Neither thermal sensors nor CCTV-based smoke detection can yet achieve this range.
There are two significant disadvantages to smoke detection. The first is that in order for a fire to create a detectable smoke column, the fire itself needs to be quite big. Big fires mean big resources are required to fight it, and in many cases, the fire is already out of control before it is detected. The second significant disadvantage is that smoke can only be detected in particular visibility conditions. Darkness, fog, smoke, rain, and wind all interfere with the ability for a smoke column to form or for it to be identified. These disadvantages apply to all types of smoke detection, whether human or automated.
A third disadvantage specific to semi-automatic smoke detection is the prevalence of false alarms. Even the most sophisticated smoke detection algorithm will find it difficult to distinguish between clouds, fog and smoke. Smoke detection solutions often require a human to manually screen all alarms in order to reduce the number of false alarms that get raised to the responder.
There are many advantages to thermal detection compared to smoke detection. First and foremost is reliability. As long as the sensor has line-of-sight to the fire, it will be able to detect its heat in daytime, at nighttime, through fog or smoke and in poor weather conditions. Furthermore the heat signature required to trigger an alert can be obtained from a much smaller fire than one required to generate a detectable smoke column, meaning the fire can be detected and suppressed much more quickly, reducing the cost of suppression and risk of the fire getting out of control.
There are two notable disadvantages to thermal detection, the first of which is the line-of-sight requirement, which can be difficult to achieve especially in very mountainous regions. The second is the limited range of thermal sensors, which is typically only a few hundred meters for accessible and affordable thermal sensors.
However, at Insight Robotics, we have a patented software algorithm that can extend the range of an affordable thermal sensor. With this algorithm we have also been able to minimise false alarms and locate the precise GPS coordinates of the fire, thereby enabling a fully automated solution.
In conclusion, neither smoke detection nor thermal detection offers a 100% guarantee of finding every fire, but advances in technology that enable us to ‘see’ fires using heat instead of smoke means we have a better chance at minimising wildfire disaster, thus protecting lives and resources.