A key component in successfully preventing widespread destruction from wildfires is early detection. In years past, fire agencies have constructed fire lookout towers staffed 24 hours a day by spotters who were relied upon to alert firefighters about new blazes in wildland areas. At the time that the first such towers were erected in the U.S. in the early 1900s, communications were accomplished via carrier pigeons and heliographs (mirrors used to flash Morse code messages), and an Osborne Fire Finder was used to plot the locations of fires.
With the advent of cell towers, satellite GPS systems, and computers, the art and science of fire detection has moved into the 21st century. This is the first of several articles that will survey the field of modern-day wildfire detection and provide some insight on how current technology enables fire agencies to better detect wildland fires.
Based in South Africa, Envirovision Solutions (EVS) fields a wildfire detection vision system called ForestWatch, which combines camera systems with computer software to determine the onset of a wildfire, map the fire line, and provide fire imagery with associated GIS information to help fire agencies deal with wildland blazes. The organization has deployed their wildfire camera systems to many remote areas in Canada and the U.S., as well as in Australia, southern Africa and Europe.
EVS camera systems can spot and accurately plot wildfires out to a distance of 15 miles in all directions, providing precise GIS coordinates to fire agencies, thus speeding their response to a fire with a minimum of delay. With near-infrared capabilities, the cameras can also spot fires through smoke, dust, or haze, further focusing emergency response on the actual heart of a wildfire.
Although there may be many vision systems on the market that could be used to spot wildfires, EVS goes one step further, providing a computer workstation with each system that uses software to detect smoke plumes in the very early stages of a wildfire. By using complex algorithms, this software can discriminate between a dust cloud, a flock of birds, or wildfire smoke.
To accomplish quick detection of wildfires, the first step in the process is to run video through an image stabilization algorithm, average and subject it to an image quality control process at the remote site by the ForestWatch ISE (Image Sampling Engine). High bandwidth links allow for the live video to be displayed at frame rates of eight images per second without any compromise to image resolution.
Once a potential fire has been spotted, the system can alert responsible parties via the internet, cell phone, pager, and landline. Multiple parties can be notified by the same site, providing redundancy in case the primary party is indisposed.
As part of an EVS installation, detailed information on the area under surveillance is included in the customized software package. As a result, the actual infrastructure can be identified in the camera views provided, known objects which might cause false alarms can be taken into account, and the system can be set up in such a way that a captured daytime photo can overlay a nighttime image to better discern what fire agency personnel are actually looking at. Names of certain features and structures can be included in the displayed image to help identify landmarks for emergency responders.
Eyes In The Dark
Many wildfires start at night, when there is a higher risk of no one detecting them or of resources being dispatched to the wrong location in the dark. Since many fire agencies do not actively patrol for fires, depending instead on calls from the general public, they are less likely to detect a wildfire that occurs in the late night or early morning hours until it grows to an unmanageable size. In fire-prone areas like Southern California, even five minutes can make all the difference between a blaze being tackled as a spot fire or turning into an $80 million monster.