Firefighters| Automated Detection of Hazards
Awards & Nominations
Firefighters has received the following awards and nominations. Way to go!
Automated Detection of Hazards
No Flames Anymore
Summary
Video: https://www.youtube.com/watch?v=GpHlXr9f1Kg&t=16s&ab_channel=AnaPowerPoint: https://drive.google.com/file/d/1WKg2EEPEBRtNWZuUYoExi8GPyTP6TTuj/view?usp=sharingIMPORTANT: If the first link doesn't work:Video: https://www.youtube.com/watch?v=X10UePd4Z0o&feature=youtu.beVideo with no music: https://www.youtube.com/watch?v=oE7lLWOSZFg&ab_channel=AnaVideo with no music: https://www.youtube.com/watch?v=fs4dijPaAA0&feature=youtu.be
How We Addressed This Challenge
Slides: 38-42, 58-59 from ppt: https://drive.google.com/file/d/1WKg2EEPEBRtNWZuUYoExi8GPyTP6TTuj/view?usp=sharing
In our project, we tried to create a program based on artificial intelligence to predict future fires.
The application developed by us aims to predict fires using a machine learning model based on the ‘Generative Adversarial Network’ framework.
To use the application, all you have to do is to enter 4 consecutive images from “NASA WorldView Snapshots” on the web interface, images that show the place of fires during those days on the earth map.
Through the interface, the images are transmitted from the client to the server where they are preprocessed, the program will generate a 5th image in which it will try to approximate the places where new fires may occur.
The web interface is made in React, the backend is made in nodejs.
The server controls the artificial intelligence module made in Python by taking the images from the client, passing them to the Python module and transmitting the results back to the client.
How We Developed This Project
Slides: 38-59 from ppt: https://drive.google.com/file/d/1WKg2EEPEBRtNWZuUYoExi8GPyTP6TTuj/view?usp=sharing
Tools: Tensorflow, React, NodeJS
PL: Python, JS, CSS, HTML, Bash
Hardware (for training): nvidia geforce 1050
Problems: A BIG problem was OOM (out-of-memory) during training, which we managed to get over after a long period of time of working with image processing and hyperparameters tuning
How We Used Space Agency Data in This Project
Slides: 43-56 from ppt: https://drive.google.com/file/d/1WKg2EEPEBRtNWZuUYoExi8GPyTP6TTuj/view?usp=sharing
To use the application, all you have to do is to enter 4 consecutive images from “NASA WorldView Snapshots” on the web interface, images that show the place of fires during those days on the earth map.
Example: https://wvs.earthdata.nasa.gov/?LAYERS=MODIS_Terra_CorrectedReflectance_Bands721,MODIS_Terra_Thermal_Anomalies_Day,Coastlines&CRS=EPSG:4326&TIME=2020-09-12&COORDINATES=-90.000000,-180.000000,90.000000,180.000000&FORMAT=image/jpeg&AUTOSCALE=FALSE&RESOLUTION=10km
Project Demo
https://www.youtube.com/watch?v=ffMvv9r2oDU
Data & Resources
NASA Worldview Snapshots:https://wvs.earthdata.nasa.gov/?LAYERS=MODIS_Terra_CorrectedReflectance_Bands721,MODIS_Terra_Thermal_Anomalies_Day,Coastlines&CRS=EPSG:4326&TIME=2020-09-12&COORDINATES=-90.000000,-180.000000,90.000000,180.000000&FORMAT=image/jpeg&AUTOSCALE=FALSE&RESOLUTION=10km