Author(s):
Christopher J Crosby* - Arizona State University
J Ramon Arrowsmith - Arizona State University
Efrat Jaeger-Frank - San Diego Supercomputer Center
Viswanath Nandigam - San Diego Supercomputer Center
Han Suk Kim - University of California, San Diego
Jeffrey Conner - Arizona State University
Ashraf Memon - San Diego Supercomputer Center
Chaitan Baru - San Diego Supercomputer Center
Abstract:
Digital data acquisition technologies such as LiDAR (Light Distance And Ranging) topography have resulted in an increase in the volume and complexity of scientific data that must be efficiently managed, distributed and processed in order for it to be of use to the scientific community. Capable of generating digital elevation models (DEMs) more than an order of magnitude more accurate than those currently available, LiDAR data offers the opportunity to study earth surface processes at resolutions not previously possible yet essential for their appropriate representation.
Unfortunately, access to these datasets for the average user is difficult because of the massive volumes of data generated by LiDAR. The distribution and processing of large LiDAR datasets, which frequently exceed billions of data-points, challenge internet-based data distribution systems and readily available desktop software.
Our approach to the distribution and processing of LiDAR data capitalizes on cyberinfrastructure developed by the GEON project (http://www.geongrid.org) to harness distributed computing resources. We utilize a workflow-based solution, the GEON LiDAR Workflow (GLW), which begins with user-defined selection of a subset of point data and ends with download and visualization of DEMs and derived products. Users perform point cloud data selection, interactive DEM generation and analysis, and visualization all from an internet-based portal. Users may experiment with DEM resolution and DEM generation algorithms so as to optimize terrain models for their application. By using cyberinfrastructure resources, this approach allows users to carry out computationally intensive LiDAR data processing without having appropriate resources locally.
