Detecting Convective Clouds in Geostationary Satellite data with Convolutional Neural Networks

14:45-15:00, January 27 @ 4A

Talk/ Overview

Convective clouds can cause serious problems for aircraft landing. As such it is important to detect them reliably and have a number of independent models. Existing techniques rely on ground based radar and lightning detection. Here I will present a new methodology based on semantic segmentation of images from a geostationary satellite called Meteosat Second Generation.

Talk/ Speakers

William Clemens

Machine Learning Scientist, dida

Talk/ Slides

Download the slides for this talk.Download ( PDF, 3279.29 MB)

AMLD / Global partners