HPC² Events

American Society of Photogrammetry & Remote MSU Chapter Seminar: Mining Earth Observation Imagery Big Data: A Computational Perspective

When:April 21-25, 2017

Where:Hillbin Hall, Room 152, 2:30-3:30 p.m.

With the advancement of remote sensing technologies, petabytes of imagery data have been collected over the past four decades covering the entire earth surface. Mining earth observation imagery big data provides unique opportunities in addressing grand societal challenges such as national water forecasting and natural disaster management. However, it also poses significant computational challenges due to the large data volume, as well as unique characteristics of spatial data including spatial autocorrelation, anisotropy, heterogeneity, and multiple scales and resolutions. In this talk, I will introduce our works on novel spatial classification algorithms (e.g., spatial decision trees, spatial ensemble) that address some of the above challenges. We will also discuss some of our ongoing research activities in applications such as flood extent mapping and tornado damage analysis.