Water Level Monitoring

Motivation

Water sources play a crucial role in sustaining the human population. To effectively protect freshwater basins from drying up, it is essential to first identify the most vulnerable ones.

Task

Image Segmentation. Predict from a time series of satellite imagery (e.g.; Sentinel-3 or Sentinel-2) images the level of water after a certain amount of time

Sources

https://global-surface-water.appspot.com/#data

Hearth Quake Monitoring

Motivation

Earthquakes are frequent and devastating natural disasters that can disrupt communications. Responding promptly after such an event is of utmost importance in order to save human lives. This can be achieved by continuously monitoring and identifying the areas impacted by the earthquake.

Task

1. The objective is to detect changes from satellite imagery using an image before the event and one after the earthquake.
2. The objective is to regress the position of earthquakes
3. The objective is to design an efficient network to deal with the previously explored tasks
4. The objective is to exploit domain information effectively

Previous Work

Quakeset

DQNC2S-V2

Motivation

Humanitarian crises are challenging to manage due to the fragmentation of information. Social media and other textual data sources have proven effective in understanding complex events. Operators need clear information from large text streams to better plan the aids. In our work, we created an online text retriever to select relevant texts for operators to answer their needs.