HR: 0800h
AN: G21C-1286    [Abstracts]
TI: Feature Extraction from Airborne Laser Scanning Measurements
AU: * Zhang, K
EM: zhangk@fiu.edu
AF: International Hurricane Research Center & Department of Environmental Studies, 11200 SW 8th Street, Florida International University, University Park Campus, Miami, FL 33199 United States
AB: Airborne laser scanning technology allows rapid and inexpensive measurements of the earth surface for a large area. Airborne laser scanning systems return a 3-dimensional (3D) cloud of point measurements from reflective objects scanned by the laser beneath the flight path. Essential GIS data such as high-resolution digital terrain models, building footprints, three-dimensional building models (DTMs), tree heights and tree density can be generated from airborne laser scanning measurements. The critical step to derive these GIS data is to develop automated algorithms to classify the millions of point measurements involved with an airborne laser scanning data set. A framework including a series of algorithms has been developed to extract features from airborne laser scanning measurements. The framework includes three major components: (1) filtering algorithms for separating the ground and non-ground measurements, (2) a region growing algorithm based on a local plane-fitting technique for segmenting building and tree measurements, (3) algorithms for deriving building footprints and simple 3D building models. Data sets from urbanized areas including the terrain, vegetation, and large commercial and small residential buildings were employed to test the proposed framework. Both qualitative and quantitative error analyses showed that the proposed framework separated ground and non-ground measurements and identified building footprints well.
DE: 1299 General or miscellaneous (1709)
SC: Geodesy [G]
MN: Fall Meeting 2005