This project is extending PhD research that is improving methodological implementation of a global-scale artificial intelligence model for analysing the impacts of bushfires, detecting the pre- and post-bushfire phase, tracking fires and sharing useful information with the stakeholder community to mitigate deadly impacts.
The spatial and temporal dataset will be collected from the repository using Mendeley or Kaggle platform for thermal infrared images. This computational complex dataset is then fed to the high-end computing Artificial Intelligence based architecture, analysing different algorithms and developing sensors. The dataset will leverage different state-of-the-art Deep Leaning (DL) algorithms by providing a global platform for training, validation and testing the thermal infrared images to further analyse the impact of Bushfire fire on land, ecosystem and communities.
Low power IoT based sensors can be deployed to detect the spread of these fires before it spreads to a wider area during pre-phase detection. Successful completion of this project will lead to unprecedented opportunities to reduce the impact of bushfires on air quality, water, land surface changes and ecology and provides future insight to identify the risk of landslide and communicate with the stakeholder and public.
To address this problem, the Airborne Sensor Facility can provide the instrument for the acquisition to capture thermal images dataset and the complex computation can than be performed in the high-end computation platform. For the said purpose, different state-of-the-art DL algorithms can be trained over the dataset to further enhance the detection accuracy and reduce the false alarms. This will save the infrastructure and communities by timely detecting the affected bushfire area and notify the emergency relief team, thus saving from huge economic loss.