DetIEEEctados| Automated Detection of Hazards

Awards & Nominations

DetIEEEctados has received the following awards and nominations. Way to go!

Global Nominee

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.

Dust of Change

Summary

Dust of Change is a website that contains information about the Saharan Air Layer (SAL), such as the fertilizing role of the minerals contained in the dust, alongside an interactive map where the user can consult for monitoring. The map uses a convolutional neural network for image recognition of the SAL, and was trained based on the High Latitude Dust (HLD) dataset provided by NASA. The automated detection of the atmospheric phenomenon allows an early alert to the population at areas of risk of air pollution brought by the dust cloud. That way, people can prepare and follow the recommended guidelines, such as staying indoors and using masks.

How We Addressed This Challenge

Dust of change

How We Developed This Project

Inspiration

To sum up with an already impaired, by COVID-19, healthcare, and economy, in June 2020, a big dust cloud stormed through the Caribbean region causing a great impact on local people's life. The plume, which appeared to stretch about 7,500 kilometers, was in the media spotlight after it reached the USA.


Our solution

A machine learning model that could work with the data we had and recognize the phenomenon.

  1. For the training model, we used, Keras interface from the TensorFlow library;
  2. For image manipulation, auxiliary libraries like Numpy, Rasterio, OpenCV, Imgaug, Matplot, PIL, and Glob;
  3. An image segmentation model with a U Net architecture;
  4. And an Nvidia GeForce GTX 1060 GPU to run the training process.

Our biggest hurdle was learning how to use and adapt it to our goal. Also, some libraries incompatibility issues, which gave us quite a headache. Nevertheless, we are very proud of our achievements, and having it working for real gave us great joy.

How We Used Space Agency Data in This Project

NASA data

We used the labeled database available on the "Phenomena Detection Challenge Resources", more specifically the High Latitude Dust (HLD), from the NASA GitHub repository, to train the Neural Network. As well as the Dust Analysis from MERRA-2 and Aqua/AIRS satellites to give us some insights and knowledge about the problem. Also, we tried to use the data from AOD reading in the Air Quality Database, but to our disappointment, the readings were in the USA only.

Project Demo
Project Code
Data & Resources
  1. https://github.com/nasa/spaceapps-phenomena_detection
  2. https://doi.org/10.1016/j.aeolia.2016.11.002
  3. https://www.scielosp.org/article/ssm/content/raw/?resource_ssm_path=%2Fmedia%2Fassets%2Fmedicc%2Fv18n3%2F1555-7960-medicc-18-03-0032.pdf
  4. https://www.bbc.com/news/world-latin-america-53177693
  5. https://www.nasa.gov/feature/goddard/2020/nasa-noaa-s-suomi-npp-satellite-analyzes-saharan-dust-aerosol-blanket/
  6. https://worldview.earthdata.nasa.gov/?v=-210.9742435153339,-104.89411792863983,181.49659384034828,96.17171105511737&t=2019-02-02-T00%3A00%3A00Z&l=Coastlines,MERRA2_Dust_Surface_Mass_Concentration_Monthly&tr=dust_storms_overview_2019
  7. https://worldview.earthdata.nasa.gov/?v=-210.9742435153339,-104.89411792863983,181.49659384034828,96.17171105511737&t=2020-06-24-T00%3A00%3A00Z&l=AIRS_L2_Dust_Score_Night,AIRS_L2_Dust_Score_Day,Coastlines&tr=dust_storms_overview_2019
  8. http://revista.domhelder.edu.br/index.php/congressodireitoambiental/article/view/1378
Tags
#artificial intelligence
Judging
This project was submitted for consideration during the Space Apps Judging process.