Team: 17
School: New Mexico School for the Arts
Area of Science: Environmental Science
Interim: Problem Definition
Forest fires have been steadily increasing in the United States. According to research from NCEI.Monitoring.Info@noaa.gov, In 2021 the US experienced 58,733 wildfires, with 7,227,371 acres of land burned. And in 2022, the number of fires has risen to 66,255, ruining 7,534,403 acres of land . Once land has been lit on fire the area's habitat and vegetation becomes unusable, the quality has fallen. People are forced to leave their homes, but they are still at high risk of being injured. Air quality is affected by the fumes, shutting down tourism and indirectly affecting the economy. Even when the fire is stopped there are still many problems; clean up, rebuilding of homes, and agricultural and soil support. After an area has burned it is most likely to experience flash floods. Because fires spread rapidly it is difficult for firefighters to track and predict where the fire will spread to, this means that even if they are able to get to the location of the fire, the fire could be spreading in a new unexpected direction making it harder to maintain and address the fire.
Problem Solution
To help avoid the damage caused by wildfires, we plan to help First Responders locate and predict the location of Fires. We plan to accomplish this by mapping and predicting a fire’s location from remote camera imagery. Using images from previous fire data, we can triangulate the location of a smoke plume. With the help of crowd-sourcing, it would be possible for people to take photos of current fires and send them directly to firefighters to help them locate and map fires.
Progress To Date
We have done a substantial amount of research about fires and triangulation. With the help of our mentor Stephen Guerin we were able to acquire images and data from previous fires. Currently we are working on writing code that will help us to triangulate smoke clouds from images. We are also looking at existing code to help find correlating points in different images of the same plume.
Anticipated Results
We hope to help accurately locate and predict where smoke plumes are heading. We anticipate being able to accurately triangulate the smoke plume, to help predict where the fire is and where it is heading. We will be using data from previous fires, which we know the location of, and compare it to results acquired from our model. After refining our program we hope to make our code compatible with a cell phone, to help make crowd-sourcing possible.
Sources:
https://inspirit.github.io/jsfeat/#features2d
https://www.ncei.noaa.gov/access/monitoring/monthly-report/fire/202213
https://www.ncei.noaa.gov/access/monitoring/monthly-report/fire/202113
https://www.alertwildfire.org
Mentor: Stephen Guerin
Team Members:
Elisea Jackson
Deisy Jaramillo
Tania Pizano
Fatima Esparza Torres
Sponsoring Teacher: Sarah Rowe