Machine Learning Applications


Machine learning is a fast-paced emerging field that is utilized frequently in our daily lives, including google search engines, self-driving cars, natural language processing and much more. Machine learning methods excel at discovering patterns and recognizing similarities within a massively multivariate data system or one composed of thousands of variables, which is a great fit for hyperspectral and multi-spectral remote-sensing satellite imagery.

We work on the development of machine learning based algorithms to retrieve aerosol and cloud properties from remote sensing measurements: see our recent paper, which is a part of the NASA NeMO-Net project:

Segal-Rozenhaimer, M., et al. (2020), Cloud detection algorithm for multi-modal satellite imagery using convolutional neural-networks (CNN), Remote Sensing of Environment, 237, 111446, doi:10.1016/j.rse.2019.111446.

Cloud and cloud shadow masks developed for WV-2 satellite based on CNN (Segal-Rozenhaimer et al., 2020)



Developing an algorithm to detect marine stratocumulus cloud types (Marine Cellular convection) from geostationary satellites to better understand their dynamics and formation mechanisms (NASA ACMAP funded project)

Time-series of SEVIRI satellite images over the South-East Atlantic Ocean, with their corresponding cloud type masks