Presentation by Dr. Xue Liu
View the slides (PDF). Listen to the audio (MP3).
Abstract: A large number of Earth observation satellites from different countries provide huge amount of remotely sensed data every day. It is estimated that a Petabyte level of remotely sensed data are being collected per day over the world, including data from either open or commercial satellites, which has led to a “big geospatial data” issue. These data sets are collected in different wavelength regions, at different spatial, temporal, and radiometric resolutions, and have been successfully used for various applications such as precision agriculture, sustainable urban management, natural hazard monitoring, climate change mitigation, and other decision support activities. The open cloud platform Google Earth Engine provides easy access to a lot of important open Earth observation data sets from Landsat, Sentinel 1, Sentinel 2, MODIS, and other geospatial data sets, and an API for processing these Earth observation and geospatial data sets. Here we mainly present results from some of our projects which combine the big data analysis capability of Earth Engine and machine learning methods, including high resolution mapping of mangrove forest cover in tropical Africa with Sentinel 1 and Sentinel 2 data, urban area mapping for CONUS with VIIRS and MODIS data, and cropland and grassland mapping for African countries using Landsat and MODIS data.
About the speaker: Dr. Xue Liu is an environmental remote sensing scientist from the Center for International Earth Science Information Network (CIESIN), Earth Institute, Columbia University, with over 20 years of experience in remote sensing, geospatial information science, and their applications. He has led or participated in many research projects related to environmental remote sensing and geospatial decision support funded by NASA, NSF/USDA, Bill & Melinda Gates Foundation, and other organizations. He has published over 30 research papers and is an IEEE Senior Member.