11:30am - 12 noon
Tuesday 2 October 2018
Artificial intelligence for automating space data processing
The major improvements in space instrumentation and telecommunications recently led to an massive increase of the available high-quality space data.
This event has passed
- Dr Panagiotis Sidiropoulos
The major improvements in space instrumentation and telecommunications recently led to an massive increase of the available high-quality space data. Indicatively, (1) the total amount of data acquired from two Viking Orbiter missions over a four-year period during the 1970s are acquired daily by a single instrument in the recent Mars Reconnaissance Orbiter mission (2) JASMIN/CEMS facility that provides in the UK data curation for environmental and meteorological datasets is currently being update to reach 40Pb storage. The space data volume explosion has a potential to significantly contribute to scientific progress, while creating a number of opportunities for new and disruptive services and products.
However, for the time being most of the new space-related technologies envisioned by both AI scientists and enthusiasts haven’t materialised. Copying from a very recent official report, ”there has been a lack of awareness and appreciation from both the AI and space industry of each other’s maturation and capability.”.
This talk, based on experience on both earth observation and deep space (Mars and the Moon) data science projects will examine the current status of AI applications for space data processing, discuss some major obstacles that the community needs to overcome and suggest directions for the use of AI to achieve fast, accurate and reliable automation of space data processing.
Dr Panagiotis Sidiropoulos is a senior research scientist in Cortexica Vision Systems Ltd. and a honorary lecturer in the Space and Climate Physics Department of University College London (UCL). He is currently working on developing AI software for on-board satellite data processing, as well as participating in R&D related to remote sensing, public space surveillance, image/video retrieval, etc.
Moreover, he is working towards establishing collaborations between Cortexica and UK academic institutions through projects, PhD students, KTPs, internships, etc. In total he is working for 12 years in industry as well as academia, during which he co-authored 15 journal articles, 2 book chapters and 93 conference publications and abstracts, while he has participated in several meetings, conferences, and workshops organised by IEEE, ACM, NASA, ESA, etc.