Hazard Detectives| Automated Detection of Hazards

Automated Detection of Hazards

Countless phenomena such as floods, fires, and algae blooms routinely impact ecosystems, economies, and human safety. Your challenge is to use satellite data to create a machine learning model that detects a specific phenomenon and build an interface that not only displays the detected phenomenon, but also layers it alongside ancillary data to help researchers and decision-makers better understand its impacts and scope.

The COVID-19 hurricane tool

Summary

I have developed a simple web application that highlights point of high risk for the population on the map.1)The application can detect point of high clouds concentration from images gained through NASA satellite VIIRS to predict imminent flood hazard.2)The application can detect point of high COVID-19 number of infections.3)It can then combines the two detections and set an alarm of high risk for the population in case of double detection

How We Addressed This Challenge

For generations mankinds was challenged by natural hazards such as flood, fires, pandemic and hurricanes. Now it’s our time to challenge this kind of phenomenons and to build something that can help us in this fight. I aim at build a simple interface that displays data about past hazards and also help governments and decisions-makers to take actions in critical moments. During the outbreak of Covid-19 pandemic I have recognized the importance of combining automatically dataset from NASA satellite about clouds concentration (that could led to flood and disasters) and dataset of COVID-19 infection numbers. I have developed a web tool that can automatically send an alarm on a map if there is a geographical location (identified by its coordinates) where the risk of double hazard (from COVID-19 and rainfall) is high. In this way governments or decision-makers can take better and more timely action.

How We Developed This Project

I were inspired by the possibility to test our machine learning abilities and to take some action in the contest of the pandemic outbreak. First we have explored nasa dataset through Leandro Camacho tutorial and take satellite maps of Nord America (SuomiNPP-VIIRS with cloud layer) in NASA worldview. The maps shows various colors. Thick ice and snow appear vivid sky blue, while small ice crystals in high-level clouds will also appear blueish, and water clouds will appear white. Then we explored the web and find Nord America maps about COVID-19 distribution of the number of infections from CDC (Centers for disease control and prevention). In these maps the more reddish the spots are, the bigger number of infections is present. My web application detects high density of red and white color and gives a couple of coordinates as output for each maps. Then it confronts the couples of coordinates. In case of coincidence of the couple of coordinates it sets an alarm on the location on the map. I had a lot of problems along the way for the gap that the online event created since I knew nobody to work with and wanted to find team members on the online platform but it was not easy. I wanted a team to work with but end up relying on my own force alone. This is just the idea of the web tool. There is no actual code.

How We Used Space Agency Data in This Project

I have made intense use of NASA space apps botcamp on youtube and Leandro Camacho video and resources on github. I have accessed unlabeled satellite immages from NASA worldview through the partial labelling provided by Global Immagery Browse Service. I have used

Corrected Reflectance (Bands M11-I2-I1). I have taken a snapshot per day of the last 14 days in png format.

Project Demo
Data & Resources

Data & Resources

Data Sources



Literature


Tags
#machinelearning #VIIRS #covid-19 #webtool #hazard #prevenction
Judging
This project was submitted for consideration during the Space Apps Judging process.