AI offers greater protection from wildfires

Regarding the environment, wildfires can cause significant damage to ecosystems, destroying flora and fauna. Forests play a vital role in absorbing carbon dioxide (CO₂) from the atmosphere, and when trees burn, they release stored CO₂ back into the atmosphere, contributing to global warming.

Wildfires can also negatively affect water quality by increasing sediment and pollutant runoff into water bodies, and large wildfires can lead to secondary hazards such as landslides and erosion, especially in hilly or mountainous areas. And then there are the economic costs; fighting large wildfires and recovering from their damage is an expensive business.

Early detection and rapid action are essential to mitigate the diverse and significant impacts of wildfires on the environment and human health, and Artificial Intelligence (AI) is making significant strides in wildfire detection. By integrating AI, we can greatly enhance our ability to detect and manage wildfires, thereby protecting our forests, ecosystems, infrastructure, and human lives.

To effectively combat wildfires with technology, we need a multi-faceted approach. Predicting where wildfires are likely to occur can help in pre-positioning resources and raising awareness about heightened risks. Rapid detection of fires allows for swift intervention, increasing the chances of extinguishing them before they spread uncontrollably. Technology could enhance efficiency or potentially automate responses to wildfires in the future.

Predicting wildfires with AI assistance

While pinpointing the exact start of a wildfire is difficult, estimating fire risk can provide reasonable predictive accuracy. Currently, fire risk is primarily assessed using weather data from satellites, complemented by information from local weather stations where available. More sophisticated models incorporate the vapour pressure-deficit (VPD), which measures the difference between actual moisture in the air and the maximum moisture it can hold. High VPD levels indicate dry conditions that can fuel large, severe wildfires.

AI can automate and enhance these calculations by integrating diverse data sources like satellite images, local sensors, and weather stations to map fire risk on a finer scale. Adding data on soil and air moisture, measured by sensors embedded in forests, could further refine these predictions by accounting for microclimates. Future advancements might even allow us to measure fuel moisture (e.g., grass and needles) more precisely, improving our predictive capabilities.

AI-based wildfire detection uses machine learning algorithms to analyse data from various sensors, identifying the presence of fires. Unlike traditional methods that depend on human observation, AI can process vast amounts of data and detect patterns indicative of fire. Machine learning models trained on extensive datasets of fire and non-fire scenarios can accurately identify fires based on smoke characteristics and other factors while minimising false positives.

Camera detection

Traditionally, lookout towers monitored by humans were used to detect wildfires. Today, cameras on these towers, combined with AI and image recognition technology, can identify smoke or fire plumes. AI algorithms enhance these systems by reducing false positives caused by dust, fog, or other visual obstructions.

For example, agricultural machinery can throw up dust when fields are ploughed, and even wind farms can be mistaken for smoke plumes from a long distance. Weather conditions, such as haze or fog, can also make it difficult for cameras to accurately identify smoke, and the time of day, particularly dawn, dusk, and nighttime, can affect the visibility of smoke in images.

By continuing to improve machine learning algorithms with more data, AI-enabled camera detection can reduce false positives and improve the accuracy of smoke detection. However, a key restriction remains that cameras typically cannot see what's happening under the tree canopy and only detect smoke plumes once they are rising above the tree canopy.

This is an important limitation as most human-induced fires start at the forest floor and smoke only breaches the canopy once the fire underneath has already grown quite large. The process can take up to several hours from ignition, in particular if the fire starts as a smouldering fire, e.g. as a result of a discarded cigarette. The delay in detection can mean that by the time fire crews arrive on the scene, they are facing a dangerous job trying to contain the fire. While infrared technology could help to complement the shortcomings of optical cameras, the resolution of these camera systems is typically too low to provide usable images for detecting fires at a great distance.

Gas sensors

Gas sensors, or ‘digital noses’, can detect fires by sensing smoke beneath the canopy layer, where fires often start. These small wireless devices send alerts upon detecting smoke, allowing for quicker response times. AI models, trained on data from both fire and non-fire scenarios, help these sensors accurately identify smoke while minimising false alarms. Continuous training with diverse data, including artificial environments replicating forest fires, improves the models' accuracy and resilience. The more diverse the training data, the better the AI becomes at distinguishing between real fires and false positives.

For example, at Dryad, we constantly feed the model data about the natural, non-fire smells of a forest as well as the smell of smoke from a burning forest from our live site in Eberswalde, near Berlin. We also collect data from our many live sites across the world where the sensors are installed. All of this data is then compiled and used to constantly improve the models before pushing out an updated version to the devices, ensuring that they are always equipped with the latest detection capabilities.

Satellites

Satellites provide a bird's-eye view, using cameras and infrared sensors to detect hotspots. AI can automate this process, but the challenge lies in resolution and update frequency. Geostationary satellites offer broad coverage but low resolution, while low-orbit satellites provide higher resolution but less frequent updates.

What satellites excel at, however, is being able to predict the development and spread of wildfires by considering various factors such as terrain, wind direction, and speed. AI and machine learning can be an immense help in predicting the development of wildfires by taking into account huge amounts of data to quickly build accurate models. This information can be passed along to firefighting and evacuation teams on the ground to help them coordinate efforts.

The future of AI in wildfire detection

There is no single solution to wildfire detection; integrating multiple AI-enhanced methods offers the best approach. Combining data from gas sensors, cameras, and satellites can provide a comprehensive view, detailing where fires start, their size, and potential spread. This integrated system would improve coordination among emergency services, ensuring resources are deployed efficiently.

Looking forward, AI could further revolutionise wildfire response with autonomous drones. These drones, guided by AI detection systems, could quickly extinguish small fires before they escalate. This vision is already being pursued by initiatives like the XPRIZE Wildfire global challenge, which aims to develop autonomous systems capable of detecting and extinguishing fires within minutes.

AI and machine learning are transforming wildfire detection and prevention. By leveraging these technologies, we can improve early detection, reduce response times, and better protect our forests and communities from the devastating effects of wildfires. With ongoing advancements, the future holds even greater promise for AI in safeguarding our environment.

Carsten Brinkschulte is CEO and co-founder of Dryad Networks, which provides ultra-early detection of wildfires as well as health and growth-monitoring of forests using solar-powered gas sensors in a large-scale IoT sensor network.