Giovanni Marchisio Seminar at the Department of Engineering Science, University of Oxford

Giovanni Marchisio (Planet Fellow and VP of Analytics)

Abstract: Satellite Earth Observations (EO), when analyzed through artificial intelligence tools, can help us measure progress towards the Sustainable Development Goals (SDGs). With the rise of distributed constellations of Earth Observation satellites, both public and commercial (Landsat, Sentinel, VIIRS, MODIS, Planet, Airbus, Maxar), there has been a general push to make earth observation data interoperable. This has led to the notion of harmonized data products that can be fused to achieve high revisit rates. The need for high cadence time series to quantitatively leverage bio-optical models of vegetation and measure the impact human activity is driven by the urgency to measure the environmental dimensions of sustainable development. As a result, new catalogues of dense temporal data will soon dominate the global EO archives. While the deluge of data makes it possible in theory to take a daily pulse of our planet, there has been little practical work on establishing foundational analytics and infrastructures that can truly leverage the temporal dimension at scale.

Under the sponsorship of the European Union’s Horizon 2020 programme, we are exploiting the notion of a virtual constellation for the purpose of updating and maintaining the CORINE (CLC) land cover product, which is the flagship of the Copernicus Land Monitoring Service (CLMS). In this context, we have fused global daily imagery from the Planet with Landsat 8, Sentinel-2, VIIRS and MODIS to create the most complete and dense, labelled, spatiotemporal, analysis-ready training sets for modelling the dynamics of land covers, combining the higher spectral resolution of Sentinel-2 (12 spectral bands) with the higher spatial (3 m) and temporal resolution (daily observations) of cloud-free, harmonized Planet Fusion imagery at 500,000 patch locations, sampled to account for CORINE land cover classes and spatial distribution, as well as country representation, across the entire territory of the EU. To stimulate the development of foundation models for Earth monitoring we plan to open source this corpus in June 2022.

Bio: Giovanni Marchisio is Vice President of Analytics at Planet where he is contributing to its mission to image the entire Earth every day and make global change visible, accessible, and actionable. He has over 25 years of experience in artificial intelligence, computer vision and data science. Previously he was Director of Product Development at Maxar/DigitalGlobe, another leading aerospace company, and CTO and VP of Engineering at Insightful Corporation, a leader in data mining and predictive analytics. He has been a Principal Investigator on 14 R&D contracts with NASA, DARPA, the USAF, the US Army and the NIH, and led a 7 year Artificial Intelligence project with NASA’s Goddard Space Flight Center. He has articulated novel ideas and architectures in the areas of multimedia retrieval, multispectral satellite image mining, and search engines.  He is the author of several papers, book chapters, the primary inventor on sixteen US patents, has been an invited speaker at the US House of Representatives, and a visiting professor with the Department of Geophysics and Astronomy at the University of British Columbia (UBC), Vancouver, Canada. He holds a B.A.Sc. in Engineering from UBC and a Ph.D. in Geophysics and Planetary Physics from UCSD/Scripps.