ArcredX| Spot That Fire V3.0
Spot That Fire V3.0
ArcredX
Summary
Our end-to-end solution is a web-application designed to inform people of nearby active fires with the help of simple visualizations powered by Augmented reality and Machine learning.
How We Addressed This Challenge
We have designed this application keeping everyone’s comfort in mind and is primarily targeted towards the people staying in the vicinity of fire-prone areas. In the present-day, information is available through a plethora of platforms. Information about a fire is mostly obtained through the local news channels, social media or maybe word of mouth. Although these sources together form a trustworthy channel, there should be an established line of communication for the people to receive verified real-time information. We hope to bridge that gap through this app. We bring in a platform for everyone to be aware of and report any fire hazard, and be able to visualize satellite data live on their mobile phones.
For detailed description, please check our github readme - https://github.com/s-ai-kia/nasa_stf
How We Developed This Project
Humanity has been facing fire-related calamities since the dawn of time and there arises challenging situations of survival even today. Wouldn't it be great, if everyone knew beforehand about any fire disasters, wildfires so that they be prepared and save their life and businesses. We being undergraduate engineering students feel the need to work towards solutions tackling major problems in society and towards technological advancement for humanity.
We have created an end-to-end solution to tackle the problem of wildfire detection, prediction and assessment. We have provided a web application built in Flask and powered by Data Analysis, Artificial Intelligence and Augmented Reality, and hosted on Heroku - https://arcredxnasa.herokuapp.com/
Our Solution:
We Analyze 4 kinds of Data:
- Satellite Data
- Crowdsourced Data
- [Future Scope] Drone Data
- [Future Scope] Sensor Data
We analyze them with Python 3 to give meaningful insights and predictions for wildfires. Please read our github readme for detailed insights - https://github.com/s-ai-kia/nasa_stf . We have discussed in depth on Prediction and Risk Factor Analysis.
The map visualizes all the fire hazards 🔴 in real time with predicted spread directions 🟡, generates risk factor predictions and severity of current wildfire. If subscribed, our platform generates alerts for users within a certain range of the origin of the fire.
We give insights in the map on crowded streets, risk score etc. We also embed Augmented Reality with AR.js into our solution to help visualize directions better when people are in a state of panic (often in the case of wildfire).
Experience our platform LIVE at - https://arcredxnasa.herokuapp.com/
Features:
- Interactive Satellite Data Analysis - https://arcredxnasa.herokuapp.com/inform
- Interactive Map - https://arcredxnasa.herokuapp.com/map
- Interactive Google map for Analysis - https://arcredxnasa.herokuapp.com/draw
- Augmented Reality - Location Aware - https://arcredxnasa.herokuapp.com/augment
- Fire Reporting Platform secured by CNN - https://arcredxnasa.herokuapp.com/
Please Check Our Platform Overview to get an all round insight - https://youtu.be/rAXw9tIKfM8
We implemented Augmented Reality for visual direction for people in stress and panic incase of a calamity. Please check our Augmented Reality here - https://arcredxnasa.herokuapp.com/
Please Check our Gitub for full details - https://github.com/s-ai-kia/nasa_stf
How We Used Space Agency Data in This Project
Satellite Data Used:
- Fire Information for Resource Management System (FIRMS)
- FIRMS - MODIS 1km (7 day) and VIIRS 375m / S-NPP & / NOAA-20 (24 h)
- CSA MOPITT Data
- NOAA Storms Event Location Data
First, we used the FIRMS MODIS 1km 7-day data for 2-D map visualization and the VIIRS 375m / NOAA-20 24-hour data for 3-D visualization on a globe. Then we incorporated the MOPITT Carbon-Monoxide content data for wildfire spread prediction. We also used the NOAA Storm Events data for wildfire point-of-origin prediction.
Project Demo
https://docs.google.com/presentation/d/1JGik0L8JIbugH7eYyxHINeCGkCFVy5zufUr99yuAMFI/edit?usp=sharing
Project Code
Data & Resources
1) MODIS Collection 6 NRT Hotspot / Active Fire Detections MCD14DL. Available on-line [https://earthdata.nasa.gov/firms]. doi: 10.5067/FIRMS/MODIS/MCD14DL.NRT.006
2) NRT VIIRS 375 m Active Fire product VJ114IMGTDL_NRT. Available on-line [https://earthdata.nasa.gov/firms]. doi: 10.5067/FIRMS/VIIRS/VJ114IMGT_NRT.002
3) NASA/LARC/SD/ASDC. (2000). MOPITT CO gridded daily means (Near and Thermal Infrared Radiances) V008 [Data set]. NASA Langley Atmospheric Science Data Center DAAC. Retrieved from https://doi.org/10.5067/TERRA/MOPITT/MOP03J_L3.008
4) National Oceanic and Atmospheric Administration
5) Imagga (Application Programming Interface) with VGG16
6) Google Maps V3 (Application Programming Interface)
7) OpenStreetMap, OpenWeatherMap (Application Programming Interface)
8) Heroku (Cloud Application Platform) to deploy our Web Application
9) Visualization: Plotly, Leaflet.js, chart_studio, Open Street Map & Google Maps
10) Machine Learning: Python 3, Pandas, scikit-learn, TensorFlow, Keras
11) Augmented Reality: AR.js
12) Design: Adobe Photoshop, Adobe Illustrator, Adobe Premiere