Exploring Data Mining Techniques for Tree Species Classification from Aerial LiDAR and Hyperspectral Imagery
Presented by Julia Marrs
Thursday March 9, 2017 @ 6:00 pm - 8:00 pm
(Eastern Time (US & Canada), Bogota, Lima)
The use of LiDAR techniques for recording and analyzing tree and forest structural variables shows strong promise for improving established hyperspectral-based tree species classifications, but previous multi-sensoral projects have often been limited by error resulting from seasonal or flight path differences. NASA Goddard’s LiDAR, Hyperspectral, and Thermal imager is now providing co-registered data over established experimental forests in the United States. This talk will cover free, user-friendly data mining applications, and evaluate techniques for integrating, simplifying, and analyzing remotely sensed datasets. This methodology opens the possibility of more easily and effectively addressing large-scale forestry questions like tree inventories and carbon sequestration on a species-specific level. Julia Marrs graduated from the Hunter College Geography Department in 2016, and is currently pursuing a PhD in Geography from the Department of Earth & Environment at Boston University, where she is expanding on her background in plant ecology and remote sensing of temperate forests. Her current research interests include spectral and fluorescence remote sensing of vegetation for investigating ecosystem productivity and greenhouse gas flux at a range of spatial and temporal scales.