As governments and firefighters attempt to switch spending and efforts from wildfire suppression to prevention, the often overlooked detection stage is seeing a technology revolution that is making a difference.
The economic, social and human costs of wildfire are staggering and getting worse. In the US for example, annual suppression costs alone have hit nearly US$1 billion in three of the past 15 years, and some predict that number will double in the next 15 years. Accounting for the loss of assets, property and human lives and livelihoods further magnifies this number.
The risk and frequency of wildfire is increasing due to climate change and poor land management allowing fuel to accumulate. On top of that, many of these fires happen in or near the wildland urban interface (WUI) – areas where forests intersect with urban development – increasing the cost of wildfire suppression and also damage.
With the costs of wildfire suppression rising so quickly, governments regularly spend significantly more than their budgets allow, often borrowing from funding intended for fire prevention. Wildfire prevention methods including public education, prescribed burning and better land management are proving to be effective in reducing wildfire risk, particularly in WUI areas. Research shows that each dollar spent on wildfire prevention can save several dollars in wildfire suppression, yet prevention continues to be underfunded.
Even if budgets allowed for comprehensive wildfire prevention, fires will continue to be a problem, and the little discussed step between prevention and suppression is detection. Early detection is crucial to effective fire suppression. The larger a fire grows before it is detected, the more expensive it is to suppress the fire, and fires allowed to grow beyond a certain size can’t be fought at all in some cases. The fire suppression community commonly supports the idea that a fire should be detected within five to 15 minutes after ignition depending on weather conditions. In any case, once a fire starts, the single greatest determination of the success of fire suppression is detection and response time.
Methods of wildfire detection
Most parts of the world with high risk of wildfire employ some method of detection. The traditional method is manned watchtowers. Humans in towers can scan an area with up to 30km radius and identify fires by the smoke columns they create. At a distance of 13km, a human can reliably spot a fire 22m2 in size, though this method is severely limited by weather conditions such as fog, clouds, reflection by sun and objects, and time of day. Manned watchtowers can be expensive, especially in regions with high cost of employment, and humans fatigue very easily when doing visual surveillance.
There are technologies that automate wildfire detection so as to minimise the human factors:
- CCTV can reduce the number of humans required to survey an area by centralising multiple views in one area. However, laying cable can be expensive, and surveillance is still subject to human fatigue.
- Satellite imagery can be used for wildfire detection, but only if there is no cloud cover. The fire size required to identify it from satellite is at least 100m2, and the scan time is at least three hours (though this is improving). At present, satellite is better suited for monitoring remote fires or fires with low risk of becoming disasters.
- Automated smoke detection algorithms based on CCTV or a series of photos can also be used to identify fires. These algorithms look for smoke behaviour by analysing multiple frames of images taken of an area. Automated smoke detection can identify a fire 100m2 in
size at 10km distance, and in testing is shown to have a similar detection rate to manned watchtowers.
However, the most recent of these technologies was introduced 15 years ago, and there has been little innovation in wildfire detection since… until now.

Thermal detection
Just last year Insight Robotics introduced a new wildfire detection technology that enables the use of thermal sensors over long distances to effectively detect and locate wildfires. This invention, which won the Entrepreneurship of the Year Award from IBM in 2014, is the first in the world to spot an emerging wildfire as small as a single 2 sq m tree within a 5km radius.
The problem with simple and affordable thermal sensors is their limited range. A simple thermal sensor can be combined with basic software to trigger an alarm when a heat ‘signal’ above a certain intensity level is received. The heat ‘signal’ emitted by a fire at 5km distance from the sensor would degrade significantly by the time it reaches the sensor. In order to trigger an alarm, the threshold level needs to be set very low. However, nearby objects that are not necessarily on fire but are hot (for example rocks or leaves in the sun) may exceed the low threshold level and trigger a false alarm. If the threshold level is increased to reduce false alarms, distant fires will not be detected.
The latest innovation in wildfire detection includes a wildfire detection robot that uses a Geospatial Intelligence System (GIS) in combination with patented algorithms to dynamically set the threshold level of the thermal sensor based on the distance of the fire, overcoming the range limitation of a simple thermal sensor. It can identify a fire 2m2 in size at 5km distance, and 20m2 at 8km distance. The robot can calculate the distance of the fire because it works with the GIS to first identify the location of the fire, and the single robot can report the location of the fire to the response team.
The limitation of thermal detection is that the sensor must have line of sight to the fire. Smoke detection makes it possible to identify fires behind a hill that a robot would not identify.
Unlike smoke detection, thermal detection is not limited by weather conditions or time of day. These robots monitor for wildfire 24/7 and will identify fires when they are very small and manageable. Insight Robotics has more than 60 robots operating in China with 100% effectiveness in detecting wildfires per the specification to date. In Jinan City in 2014, eight robots identified 173 of 173 fires, none of which grew larger than 20m2 by the time they were suppressed.
Prevention, detection, suppression
Minimising wildfire disaster requires a comprehensive strategy for preventing, detecting and suppressing wildfire effectively. The already high and continually growing costs of suppression can be drastically reduced if prevention and detection mechanisms are properly employed. By increasing investment in prevention and early detection, overall spending will be reduced, and those funds can be used for other public and government spending. Furthermore, the impact of wildfire on assets, society and humans will also be reduced. Fire managers won’t be able to stop building in the WUI or oncoming climate change, but they will be better able to prevent disasters caused by wildfire.
For more information, go to www.insightrobotics.com
You must log in to post a comment.