FireGoApp| Spot That Fire V3.0
Awards & Nominations
FireGoApp has received the following awards and nominations. Way to go!
Spot That Fire V3.0
FireGo: detects, confirms and mitigates forest fires
Summary
With FireGoApp we seek to disseminate the analysis of the economic and ecological impact of a fire in forest areas and thus generate awareness and encourage the public support for rescue services. The detection of forest fires in early stages is done by generating alerts that reach the closest users to the possible fire, who will have the possibility to verify the alert with the confirmation from the request of sending multimedia material of the fire, avoiding false alarms and improving the learning model used. This alert, confirmed by several users, is sent to the competent authorities with special information (geographical, climatic, etc.) in order to proceed with the action.
How We Addressed This Challenge
We developed a code in Google Earth Engine that is a platform that allows us to view and analyze satellite images through datasets such as USGS Landsat 8 Collection 1 Tier 1 TOA Reflectance and FireCCI51: MODIS Fire_cci Burned Area Pixel product, version 5.1, thus, obtaining one of the potential features to feed a neural network that allows us to detect forest fires in early stages. With the data thrown by the algorithm, the aim is to supply the information to FireGoApp so that a user receives alerts about possible nearby fires and confirms their existence and severity so that the corresponding authorities can take action.
This is important because the detection of the fire in early stages together with the additional information provided to the competent authorities, can speed up and improve the process of fire mitigation, reducing the impact of fires on ecosystems, as well as the loss of flora and fauna and the environmental and economic impacts that these can entail.
In addition, one of the most frequent causes of environmental fires is ignorance of the consequences that this can have. This is achieved through the generation and dissemination of economic impact data to the users of the PPP (historical study of rainwater consumption and solar radiation in the affected area); and a study of the risk of collateral damage, calculating the proximity to productive crops, population centres and infrastructure), historical data (the longevity of the affected vegetation) and ecological data (a study of the gases emitted during the fire and the effect of the fire on the climate during and after its spread). This covers the part of estimation and evaluation of the impact of forest fires.
With the above, we seek to reduce the environmental and historical impact of forest fires, in which years and even centuries of natural history are lost, and ecosystems are affected, even those that are more scarce such as the moors.
How We Developed This Project
What inspired your team to choose this challenge?
In Colombia you can find páramos (among them Sumapaz, the largest in the world), beautiful ecosystems vital for maintaining water cycles, which are found in few countries in the tropics, and which are also very important for the biodiversity of fauna and flora they shelter. These moors are seriously threatened by forest fires which can destroy vegetation that has been recovering for centuries and play a fundamental role in the function of the moor. The difficulty of such fires is that they sometimes occur in areas that are difficult to access for the competent authorities and are detected long after the fire has started because of the remoteness of the sites. This also happens in many ecosystems, which is why we considered it essential to attack this problem by detecting and providing early warning of fires, with information such as their location, geographical accessibility, proximity to population centres, the type of vegetation in the area, whether it is more or less prone to burning, whether there is a body of water nearby and of what nature, among others; information that can be useful in generating an optimal plan of action.
What was your approach to developing this project?
Our main focus was technical, using satellite data and machine learning, development tools such as Earth Engine, and the development of a mobile PPP.
What tools, coding languages, hardware, software did you use to develop your project?
We used Earth Engine, a cloud-based geomatics platform that allowed us to visualise and analyse satellite images of interest, which it codes based on JavaScript. For the prototyping of the app, we used Figma, a rapid prototyping tool, and for the development of the machine learning model, we planned to code in Python.
What problems and achievements did your team have?
We had a problem with the definition of the problem we wanted to address until we finally decided on the one we had already discussed, and also with the use of the Earth Engine tool. We were able to interpret the satellite images and determine the area consumed by the fire at a given time, as well as develop a prototype of the APP and its functions.
How We Used Space Agency Data in This Project
The following datasets were used in this work:
USGS Landsat 8 Collection 1 Tier 1 TOA Reflectance
Landsat 8 Collection 1 Tier 1 calibrated top-of-atmosphere (TOA) reflectance. Calibration coefficients are extracted from the image metadata. See Chander et al. (2009) for details on the TOA computation.
FireCCI51: MODIS Fire_cci Burned Area Pixel product, version 5.1
The MODIS Fire_cci Burned Area pixel product version 5.1 (FireCCI51) is a monthly global ~250m spatial resolution dataset containing information on burned area as well as ancillary data.
These were used as raw materials to understand the magnitude of the problem in historical terms of intensity, frequency, areas affected by fires (small, medium and large) near populated areas, protected forest areas, cultivation areas, etc.
To visualise this information, use was made of Google Earth Engine, a planet-wide platform for Earth science data and analysis driven by Google's cloud infrastructure. This made it possible to process the data and propose the initial part of a machine-learning-based algorithm for feeding data into FireGoApp.
Project Demo
https://docs.google.com/presentation/d/1hIENkhZVFeG3aw142RxItIAKp48Rye2UA5P1-m_Tx6M/edit?ts=5f7a76f6#slide=id.p
App Prototype: https://www.figma.com/proto/Mu2gOK7hBBZ5aUmL3gCmir/Untitled?node-id=5%3A2&scaling=scale-down
Data & Resources
The following datasets were used in this work:
- USGS Landsat 8 Collection 1 Tier 1 TOA Reflectance
- Landsat 8 Collection 1 Tier 1 calibrated top-of-atmosphere (TOA) reflectance. Calibration coefficients are extracted from the image metadata. See Chander et al. (2009) for details on the TOA computation.
- FireCCI51: MODIS Fire_cci Burned Area Pixel product, version 5.1
- The MODIS Fire_cci Burned Area pixel product version 5.1 (FireCCI51) is a monthly global ~250m spatial resolution dataset containing information on burned area as well as ancillary data.
- https://repositorio.gestiondelriesgo.gov.co/bitstream/handle/20.500.11762/28309/Cartilla_Incendios_2019-.pdf?sequence=4&isAllowed=y