Hotshots Staff| Spot That Fire V3.0
Spot That Fire V3.0
Hotshots Staff
Summary
Hotshots Staff is an AI-based cross-platform solution that is powered by data from geostationary and polar orbit satellites. AI makes the correlation of temperature data, FRP (Fire Radiative Power), humidity, precipitation, wind speed and direction to generate indicators of fire risk, critical fire hazard and PPFI (projection of spread of fire foci). Using pre-existing mathematical-mathematical models to calculate them in real time, being responsible, ppfi for example, for enabling strategic fire fighting planning as well as a better understanding of the destroyed areas, the damage to biodiversity and the impact on the economy.
How We Addressed This Challenge
Hotshots Staff is a multiplatform AI-based solution, which is powered by data from geostationary and polar orbiting satellites that are released free of charge by federal regulatory agencies INPE and INMET. AI is called ANNIE (Artificial Neural Network Intelligence of Environment) and it correlates data on temperature, FRP (Fire Radiative Power), humidity, days without rain, wind speed and direction to generate fire risk indicators , critical fire hazard and PPFI (projection of spread of fire outbreaks). ANNIE, through historical analysis of the data made available by the satellites and a study of the interrelation between the various types of data, is able to predict the periods and places with the highest density of probability of fire occurrence. The AI goes beyond the study of environmental variables and analyzes the data from past fires with their respective impacts, and to predict places and periods that exist environmental and historical indications that there may be large fires (critical fires). Always considering the advance of the agricultural frontier as a factor that pushes the demographic occurrence of fire along with it. PPFI (fire spread propagation projection), and a real-time data that indicates the vectors of fire spread based on several variables, mainly three, wind speed and direction and FRP (Fire Radiative Power), PPFI describes the most likely trajectory of fire and determines its catastrophic potential. With the PPFI, we can understand the damage of a fire before it occurs, a fact that allows preventive action in order to avoid damage to the ecosystem and the economy, thus saving the lives of people and animals.
How We Developed This Project
After the team was formed during the week in which the event started, on Friday night we started working, defining the challenge that we would seek to solve! To better work, we made, through a video conference meeting, an alignment to better understand the challenge, what we needed to do and deliver at the end. We define the name of the team, the motto, we set the goals, we define and delegate the tasks together with their respective deadlines, so that we can meet them in time. On Saturday we woke up very early and 8 am we were hitting (almost literally) on the tool we set to work (Discord and a little Google Meet). During Saturday morning, we held simultaneous meetings to align individual tasks and start production. While one sub-team was focused on searching for statistical data to be used as a feed for artificial intelligence, the other part was focused on developing the presentation and prototyping of our solution. Early Saturday afternoon we held a general meeting for alignment in which we developed the procedural flow of our solution, defining the inputs and outputs that it would have. Throughout the rest of the day, we continued to work hard and designed our logo using the InkScape tool. Still using the same tool, we developed the application screens. For AI we developed a Python code using Visual Studio Code and Jupyter Notebook in addition to using some pre-defined libraries (Sklearn, Numpy, Pandas, Matplotlib). At the end of the project, we used github as a repository. We stayed together all night, and even now, at dawn, none of us have slept yet and we continue to work to deliver all our goals in the best possible way. After everything was ready, we developed the video pitch using the Adobe After Effects tools for compositing, Cinema 4D for 3D animation and Adobe Premiere for video and audio editing. It was a lot of work, we spent more than 24 hours in a row without sleep and with practically uninterrupted work, but we managed to reach our goal, which was to deliver what we had in mind, considering the deadline we had.
How We Used Space Agency Data in This Project
The data from NASA and partner agencies influenced the filtering of the global behavior of forest fires, helping to compose the study that united the vision through the information obtained by them (and more specifically by the AQUA satellite) with the environmental and economic impacts during and after fires helping to model the general behavior of the problem. In this way, the machine learning system created was able to predict, with few fluctuations, in the latitude and longitude information of the expected future outbreaks of fires in Brazilian cities (Cáceres - Mato Grosso, Brazil), which were hit hard this year; In addition to helping to create a prototype of an application linked to AI to help the general public to access, be notified and know the general impact (environmental and economic) about the event.
Data & Resources
Data bank of fires. [S. l.], 2 ago. 2020. Available at: http://queimadas.dgi.inpe.br/queimadas/bdqueimadas#exportar. Accessed on: 4 out. 2020.
INMET Meteorological Data BANK. [S. l.], 2 ago. 2020. Available at: http://bdmep.inmet.gov.br/. Accessed on: 4 out. 2020.
Liu, William Tse Horng. Remote sensing applications. Text Workshop, 2015.
de Moraes, Elisabete Caria. "CHAPTER 1 FUNDAMENTALS OF REMOTE SENSING." (2002).
EPIPHANIO, José Carlos Neves. "Remote sensing satellites." Course on the use of remote sensing in the study of the environment. São José dos Campos: INPE (2002): 37.
OLIVEIRA, Jean Carlos Pinto de Arruda; DA SILVA, Dércio Santos; BEZERRA VIEIRA, Flávio Gledson. REMOTE SENSING APPLIED TO FIGHTING FOREST FIRE IN SERRA DE SANTA BÁRBARA DE MATO GROSSO STATE PARK. XVIII National Fire Brigade Seminar, Scientific Magazine of the Pernambuco Military Fire Brigade, v. Vol.04, ed. No11, 23 Aug. 2018.
Zorgetto, Afonso Assalin, and Vandoir Bourscheidt. "Temperature Response to Different Soil Uses from Surface and Remote Sensing Data (Land use effects on the surface temperature evaluated by ground level and remote sensing data)." Brazilian Journal of Physical Geography 11.4 (2018): 1416-1428.
Setzer, Alberto Waingort, Raffi Agop Sismanoglu, and José Guilherme Martins dos Santos. "FIRE RISK CALCULATION METHOD OF THE INPE-VERSION 11 PROGRAM, JUNE / 2019." CEP 12: 010.
Nascimento, Diego Tarley Ferreira, and Juliana Ramalho Barros. "Identification of heat islands by means of remote sensing: a case study in the city of Goiânia – GO / 2001." (2009).
Ribeiro, Guido Assunção. "Study of fire behavior and some effects of controlled burning in stands of Eucalyptus viminalis Labill in Três Barras, Santa Catarina." (1997).
Narciso, M. G., et al. "Proposed method for selecting the fire risk indicator by region." Embrapa Pantanal-Research and Development Bulletin (INFOTECA-E) (2011).
INFOQUEIMA: Monthly monitoring bulletin. Burn program, satellite monitoring, São José dos Campos, SP, Brazil, vol. Vol.05, p. 1-11, Jan 1 2020.
NASA. 2020. Use of models for general systems modeling:. Available at: https://firms.modaps.eosdis.nasa.gov/active_fire/. Accessed in: day, month and year.
NASA. 2020. Use of models for general systems modeling:. Available at: https://www.nasa.gov/geonex/. Accessed in: day, month and year.