Big Data Analytics and Machine Learning Tools for Space and Earth Observation

Space and Earth Observation, as an area of active research and exploration, is an example of science in which information supply is experiencing steady and significant year-on-year growth, with the number of instruments and investment increasing leading to demands of speedy processing. In addition to the volume, velocity and value, the data has significant degree of uncertainty, and the actual values that are of interest require substantial computational resources if inferred by traditional methods. Dealing with this sheer volume of data at a tremendous speed cannot be achieved by human operators alone.

Moreover, “blanket” modelling approach whereby High-Performance-Clusters are employed to solve forward and inverse problems replying upon conventional machinery is nearing its capacity and is not cost-efficient. In addition and apart from technical issues, there is a growing demand from the end-users of EO work, customers and society to receive answers to queries that are only implicitly related to the outcomes of traditional physical EO modelling. Answering these customer-driven questions require embedding of sophisticated multi-scale Data Analytics, Machine Learning, and Artificial Intelligence tools in the analysis loop.

This PhD project will focus on developing Data Analytics and Machine Learning tools that are capable of responding to the challenge of Big Data in the Space and EO industries.