Inspiration
The inspiration for our project stemmed from the urgent need to address the escalating threat of wildfires in Alberta. Last year, we witnessed the devastating impact of these disasters on communities, ecosystems, and economies. We fueled our commitment to developing solutions using AI.
What it does
Our project harnesses the power of AI to predict the impact of wildfires in Alberta. It was done by integrating diverse datasets, including weather, vegetation indices, and population information. The developed AI models serve a dual purpose: identifying potential fire-prone regions and predicting the severity and extent of wildfires. This comprehensive approach provides actionable insights for effective firefighting and community protection.
How we built it
We built our solution through a meticulous process that involved fundamental analysis, data preprocessing, and the development of two distinct AI models. We utilised a range of datasets, including historical wildfire data, ERA5 weather data, MODIS satellite data, and population information. The AI models were trained and validated using a combination of these datasets, incorporating advanced techniques such as LSTM for fire probability estimation and U-net with ResNet for burned area prediction.
Challenges we ran into
Throughout the development process, we encountered challenges related to data preprocessing, model optimisation, and downloading datasets. Handling misinformation in data required careful consideration.
Accomplishments that we're proud of
We are proud to have developed and validated two AI models that address different aspects of wildfire prediction. Our models demonstrate high accuracy and incorporate real-world considerations. The comprehensive approach, considering vulnerability factors and seasonality, adds a layer of sophistication to our solution.
What we learned
Throughout this project, we learned the importance of interdisciplinary collaboration and the need for diverse datasets in developing effective wildfire prediction models. We gained insights into handling complex temporal and spatial data, developing AI models, and ensuring the practical applicability of our solutions.
What's next for Wildfire
We plan to enhance our models by incorporating additional data sources, such as weather station data and fire weather indices. Exploring the inclusion of carbon considerations and further refining the models based on ongoing research will be pivotal.
Log in or sign up for Devpost to join the conversation.