scan-Combustion| Spot That Fire V3.0
Spot That Fire V3.0
Rainformation
Summary
Through technological resources, the project allows rapid detection and combat of fire outbreaks through the integration of sensors, drones, towers and satellites. Information is processed in real time and allows authorities to respond quickly. Towers equipped with thermal sensors cover an area of approximately 7,000 football fields. The drones are equipped with CO2 meters and operate in a monitoring system that allows selecting the closest to the fire source to perform CO2 analysis. Satellites are responsible for storing this data in the cloud, while servers are dedicated to processing information.
How We Addressed This Challenge
Firefighting has become an important human action to preserve ecosystems and guarantee the quality of life for the entire population. The present project aims to enable, through technological resources, a rapid detection of possible fires even in remote regions and, through the integration between different sensors, drones, communication towers and satellites, to send information in real time with high precision to authorities allowing actions to be taken immediately. The installation of towers with a range of up to 5 miles forms a covering network capable of protecting an area of up to 78.54 km², the equivalent of more than 7,000 soccer fields. Drones have a monitoring system that allows them to move individually and that allows communication between drones, which always allows you to select the drone closest to the possible fire source based on the real time of the journey. Each drone is equipped with a CO2 meter and based on its specifications, it is possible to create a ratio of one drone for every nine communication towers.
Communication towers, in turn, they have thermal sensors with a range of up to 5 miles that monitor the temperature of the regions, in search of possible fires. These thermal sensors can rotate 360 degrees and thus cover an area of approximately 70 km². The satellites are responsible for receiving information in real time from the communication towers and then storing this data in the cloud, while specific servers will be dedicated to the processing of this information, in order to identify possible outbreaks of fire and, thus, send alerts the authorities for a quick response to avoid further environmental damage. With the use of APIs, it is also possible to classify certain regions at different levels of fire risk as a way of alerting residents and local authorities. The project addresses the challenge chosen since the purpose of the developed proposal is to detect fires prematurely before the burning spreads. Consequently, this would result in a context in which authorities would be better able to fight fires because they would deal with the problem right from the start. In view of this, the project presents an improvement in the identification process for firefighting where it is applied, thus addressing one of the expected results of this challenge proposed by NASA.
How We Developed This Project
Every year, Brazilian forests have a lot to face in terms of fighting fires. Last year, without control, the flames took over part of the Amazon and caused irreparable damage to the local fauna and flora. Recently, we also had a fire episode in the Pantanal regions, which also ended up affecting the lives of many animals - including some in extinction - and the forests. Bearing this in mind and the difficulty of containing the flames of a fire that has already spread, our team decided to select and propose solutions to this challenge because it represents a reality for all of us and for presenting serious risks to our local fauna and flora.
Our approach in solving this challenge was to propose a method of preventing possible fires in order to avoid major problems in cases of containment. Currently, prevention systems already exist, however, we are looking for an early detection approach in real time, in order to promote a containment of spread in its initial phase.
As software implementation resources, we proposed the development of a program that would make predictions based on Machine Learning and object detections - which could be applied through the Python programming language and other resources such as the OpenCV library. Among other functions attributed to the software are: control of drones in an automated way, issuing a fire alert to local authorities and mapping the towers' coverage environment for viewing.
In terms of hardware, we proposed the use of towers installed in the required regions with cameras with a 360º swiveling capacity, CO2 sensors and drones.
Our problems came down to creating a good fire detection and estimation strategy and what would be the best layout of the main structure. Achievements apply to the same topics, given that we managed to adjust them.
How We Used Space Agency Data in This Project
To make the information more accurate, Rainformation uses public information released by entities such as NASA and INPE (Brazil) to have access to the locations of fires that occurred in previous years in order to enable more accurate analysis and, thus, determine the probable causes, such as arson or natural fire. Access to data is done through the use of APIs and automatically, which allows processing a greater number of information based on machine learning to obtain more and more accurate information.