Colloquium Computer Science - Dr. Trienko Grobler, University of Stellenbosch, South Africa

Title: Application of Machine Learning in Remote Sensing and Radio Interferometry

Abstract:

Remote sensing is the science of creating information from observations of Earth made by distant platforms like satellites. Radio interferometers are instruments that consist of radio antennas that work together to observe the radio sky. In both disciplines recordings are made in different spectral bands, in the one field we observe the universe and in the other we observe the Earth. In this talk we explore a bunch of different ways Machine Learning (ML) can be applied within these two disciplines. In the case of Remote Sensing, we focus on how ML is used to perform land cover classification and change detection, while for interferometry we explore how Radio Frequency Interference (RFI) can be detected and discuss the role that other ML tasks like image classification, retrieval and anomaly detection play (or can play) in interferometric reduction pipelines.