High-Level Project Summary
Addressing recent devastating wildfires, our project aimed to create a web app that'll use satellite images to detect wildfire and predict probable pathway. As a part, we developed an algorithm capable of identifying and analyzing fire from satellite image . This algorithm is competent to determine the spread rate, frequency and predict its route by studying the weather components that influence its outburst along with mathematical formulas. Position of fire and predicted route will be shown along with the location of a user in an interactive map and also notify the user about his incoming danger. Also, relevant statistics will help govt and firefighters in pre and post management system.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Our team developed a web application capable of processing satellite images of any wildfire arising throughout the world. When a fire escapes initial attack, our team carefully assess
the overall fire situation. The purpose of the prediction process, therefore, is to enable our team to estimate what a fire will do under the expected weather and existing topographic conditions. These procedures actually form a short-term planning system that uses observations of fire behavior, fuels, topography, and weather forecasts to give advance notice of the kind of fire that can be expected.
The algorithm working beneath first identifies the region burning at that time and marks it as a red pixel. After identifying it's necessary to know the frequency and spread rate of fire. This would have been possible easily if we could gather few real-time data from that region in the very first place. But it's arguably impossible for our team to do so. As an additional plan, we gathered a few satellite images of that fire approaching from FIRMS(Fire Information for Resource Management System) earth observatory satellite images which enabled us to find out the initiation of the fire and study the nature of fire. Actually we can get the location of fire from high altitude images as they are marked with red pixels using the MODIS and VIIRS image processing system designed by NASA.
The collected image looks something like this:
Our algorithm needs some variables to be provided, such as the wind velocity and direction, terrain slope(which was predetermined for avoiding complexity), total area of the selected region, other than the selected latitudes and longitudes of the particular region and also the latitudes and longitudes of the user .
Having gathered these resources, this algorithm, processing the data, will predict the movement of the fire and illustrate it with an interactive map.
The processed image looks something like this:
longside, we will get 3 distinguished regions graded as red, yellow and gray implying usual meaning, such as the red implies that the raging fire is already going on in that region, yellow implies that the fire will already be there in the provided time and gray typically implies safe zone. Besides, the user may also have access to several maps based on the probable distance that the fire would travel in that particular time. We are using a predefined location for our models in order to better demonstrate the effectiveness of our solution. But even if a user allows sharing location, our program will detect his exact location with latitude and longitude coordinates and show his/her location on that map.
Furthermore, it will notify the user at which zone he was right that moment and advise his action. It'll be much helpful for an individual to leave the risky area vulnerable to wildfire. Through the use of this interactive, user-friendly website, users, city officials, and private organisations can not only view our model's detections and forecasts but also see the actual extent of damage done by the wildfire in that particular region, helping them to prepare for danger and reduce loss of property and lives.
The data analysis from the image looks something like this:
Moreover we are also using a using a unique twilio api, which can be used to send messages to people who are indeed in danger zone and have registered:
The Fire Prediction Process
When a fire escapes initial attack, our team carefully assess the overall fire situation. The purpose of the prediction process, therefore, is to enable our team to estimate what a fire will do under the expected weather and existing topographic conditions. These procedures actually form a short-term planning system that uses observations of fire behavior, fuels, topography, and weather forecasts to give advanced notice of the kind of fire that can be expected. Typical steps that we took in this process are as follows:
ASSESS THE PAST AND PRESENT FIRE SITUATIONS
What has the fire done before you were able to observe it and what is it doing now? In both cases, try to determine what type of fuels the fire has been burning in, and what fuel stratum has been carrying the fire. What has the weather been?How has the fire responded to the weather in terms of intensity rate of spread and direction?
SIMATE INPUTS
The greatest challenge was considering all the environmental factors, such as fuels, weather, and topography.There are a number of problems to consider:spotting and crowning,
potential fuel types and fuel maps, topography, and predictable diurnal weather conditions, wind speed and its direction.We have taken all of them into account.
CALCULATING FIRE BEHAVIOUR AND INTERPRETING OUTPUTS
Estimating all the inputs and considering weather forecasts, we calculate the fire behaviours:spread rate and area predictions.For new fire starts , fire growth as an elliptical pattern on the ground is estimated in terms of area considering time periods. If on a slope, the procedures used for predicting area also considers it . The length-to-width ratio of the ellipse is governed by the windspeed and steepness of the slope. We have considered all combinations of wind and slope, including fire burning upslope or backing downslope and with wind blowing either up, down, or cross-slope. The fire growth for a specified time period is then projected on a map.We predict this fire behaviour for associated authority and also individuals surrounding the area.
Space Agency Data
Data from NASA
https://firms.modaps.eosdis.nasa.gov/ [Fire Information for Resources Management System-FIRMS]
Hackathon Journey
The path we traversed to get to the position where our team currently stands has been quite adventurous. The journey has been filled with memories, emotions, laughter, and sweat. We as a team have always been sticking up together while taking and accomplishing various milestones. Human beings have the natural tendency to find solutions to the problems he sees and experiences with his own eyes rather than the problems he never saw or faced. From our childhood, we have always been intimidated by the raging problems we faced in our studies and always have sought a way through those tough mathematical problems. We always did, that’s basic human nature. Social problems never used to bother us when we were infants as we never had the capacity to comprehend the situation or the capability to develop a solution to those problems. This changes when we grew up.
We six have always been sticking close to social problems and their solutions for quite a while now after realizing that small steps bring about the grand change and it’s high time the youth took the responsibility of the society with modern thoughts and technologies.
As the start of the pandemic brought disaster in the lives of the people, we have been standing beside them forming an organization of our own and dedicating various fundraising and projects to help the people affected by covid and various other issues that follow. Likewise, this year the unusual increase in Wildfires caught our attention and our minds have been stuck to find a viable solution to this disaster which is currently engulfing the western countries and a threat to the environment. Sticking up to our plan we laid out a blueprint on how to progress with our thoughts on developing them into a real-world solution. Each one of us specializes in different sectors starting from Graphics design to coding, thus we always had something new to learn from each other. Starting from those late-night meetings when an idea hit our mind to exporting those illustrations, things were jampacked. The bug fixes and the errors were always there to annoy as the cherry on top. Every one of us has worked hard to put a specific charge on each of our standpoints that firmly hold on to our project, ‘Project Dabanol’. We would also like to take a moment to thank our parents who have always been supporting and inspiring us throughout this journey and also to BASIS and NASA for organizing such an amazing event. Here’s to a better future!
References
Data and Resources
https://www.fs.fed.us/rm/pubs/rmrs_rp004.pdf [Farsite]
https://www.fs.fed.us/rm/pubs_series/rmrs/gtr/rmrs_gtr371.pdf
https://www.frames.gov/documents/behaveplus/publications/Anderson_1983_INT-RP-305_ocr.pdf
https://www.fs.fed.us/rm/pubs_int/int_gtr143.pdf
Data from NASA
https://firms.modaps.eosdis.nasa.gov/ [Fire Information for Resources Management System-FIRMS]
weather forecast:
https://www.timeanddate.com/
Tags
#pixel_track #wildfire #life # save_forests #fire_track #reduce_pollution #dabanol