Dryad’s AI LoRa network detects wildfire in Lebanon in 30 minutes
An AI-enabled LoRa-based wireless system developed in Germany has detected a wildfire in Lebanon within 30 minutes, allowing
The Silvanet system developed by Dryad successfully identified and raised the alarm about an unauthorized fire at a pilot site site in Lebanon (see video below).
The Silvanet system detected the small illegal fire earlier this month, allowing for a prompt response and highlighting the effectiveness of the technology in real-world scenarios and its role in enhancing environmental safety.
- Dryad hunts the firestarters
- Dryad raises €10.5m to warn of wildfires
- Dryad adds to Series A to detect forest fires
Lebanon loses 1,500 hectares of forested areas to fires every year, posing threats to communities, natural habitats, and critical infrastructure. In June this year, a massive wildfire burned in the mountainous Sinn region of Akkar, consuming 90 hectares of forest, devastating centuries-old trees, and causing damage to several homes.
With 139,000 hectares of forest, equivalent to roughly 13% of its land area, Lebanon’s forests are increasingly at risk. Extended periods of drought pose a threat, elevating the risk of fires in higher-altitude regions that historically experienced fewer wildfires. Additionally, pest outbreaks are contributing to the desiccation of trees before the onset of the fire season. Lebanon’s renowned cedars, along with junipers and firs, are among the endangered forest ecosystems.
Local broadband operator Mada is using its telecommunications expertise to deploy Dryad’s Silvanet at a pilot site at Deir Mar Moussa in central Lebanon. The network uses artificial intelligence-enabled gas sensors within a large-scale, solar-powered mesh network embedded in forested areas.
The pilot deployment has been deployed since January 2023 with 91 sensors and 2 gateways, covering an extensive 78-hectare area in the forest adjacent to Deir Mar Moussa.
The incident involved a farmer burning dry grapevines, illegal activity that posed a severe threat as the fire could potentially spread to nearby forested areas.
The system detected a change in air composition through the Bosch BME688 gas sensor and subsequent gas scans identified hydrogen, carbon monoxide, and volatile organic compounds (VOCs). Silvanet’s AI then analyzed the patterns, predicting a 70% probability of smoke at 10:33 am, triggering an alert through the Silvanet mesh network to the customer.
Silvanet’s AI model is the result of extensive research performed over the past three years. The system’s machine learning models are finely tuned and can be adapted to the specific environment of a deployment site, minimizing unnecessary alerts and enhancing the reliability and sensitivity of wildfire detection. To reduce network load and enable large-scale deployments, the fine-tuned AI models are executed in the sensors, distributed over-the-air throughout the network without requiring physical maintenance.
Mada plans to extend the pilot to a full-scale deployment, safeguarding Mount Lebanon, a region of immense ecological relevance, extending along the entire country parallel to the Mediterranean coast. This expansion underscores the significance of proactive measures in protecting communities and natural landscapes.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
