Extracting structural vegetation components from small-footprint waveform LiDAR data in savanna ecosystems

Show full item record

Title: Extracting structural vegetation components from small-footprint waveform LiDAR data in savanna ecosystems
Author: McGlinchy, Joseph
Abstract: Research groups at Rochester Institute of Technology and Carnegie Institution for Science are focusing on characterization of savanna ecosystems using data collected from the Carnegie Airborne Observatory (CAO), which integrates high spatial resolution imaging spectroscopy and waveform light detection and ranging (LiDAR) data. This component of the larger RIT ecosystem project evaluated the extraction of waveform features from the small-footprint waveform LiDAR data and their ability to explain structural variation across differing land cover types by way of bare ground cover, woody, and herbaceous biomass estimation. The combination of nanosecond digitization of the backscattered signal and fine spatial resolution allowed for the extraction of structural information embedded within the waveform at the sub-object level. Signal processing approaches subsequently were used to combine measurements within a composite footprint size of sub-meter and above. The ability of the waveform features to estimate the level of bare ground coverage for a single pixel was verified by comparing grayscale maps of the features to imagery and abundance map combinations of spectral endmembers. The ability of above-ground waveform features to explain the woody and herbaceous field measurements was evaluated statistically by forward selection regression models, selecting features significant in the explanation of the biomass field measurements. Overall, the waveform features were able to explain 70% of the variation in woody biomass across the study area. These structural features were also able to explain more than 80% of the variation in woody biomass in two out of three land use environments sampled within the study area. On the other hand, the waveform LiDAR data were only able to explain variations in herbaceous measurements in one of the three land use environments. This was attributed to the narrow range of measurements and the senescent state of the vegetation; the field data were collected in fall season 2008. The amount of herbaceous variation explained in this area furthermore was dependent on the lower limit of the range of measurements considered, which is in turn related to the limited laser-target interactions for low biomass regions. These results indicate that small-footprint waveform LiDAR data can effectively be used as a single modality to describe heterogeneous woody cover in a savanna environment. However, further research is warranted during the full growing season to more fully evaluate its performance in describing herbaceous cover.
Record URI: http://hdl.handle.net/1850/13204
Date: 2010-11-15

Files in this item

Files Size Format View
JMcGlinchyThesis11-15-2010.pdf 25.01Mb PDF View/Open

The following license files are associated with this item:

This item appears in the following Collection(s)

Show full item record

Search RIT DML


Advanced Search

Browse