[Back]

V. Ducic, M. Hollaus, A. Ullrich, W. Wagner, T. Melzer:
"3D Vegetation Mapping and Classification Using Full-Waveform Laser Scanning";
Talk: International Workshop 3D Remote Sensing in Forestry organised by EARSeL - SIG Forestry, Univ. of Natural Resources & Applied Life Scie; 2006-02-14 - 2006-02-15; in: "Proceedings, International Workshop, 3D Remote Sensing in Forestry", T. Koukal, W. Schneider (ed.); (2006), 211 - 217.

Classifying vegetation points from 3D airborne laser scanner (ALS) point clouds is a challenge and
focus of current research. In particular, low vegetation points are very difficult to identify. The basic
problem is that so far the majority of ALS systems have provided only the 3D coordinates of scattering
objects and most of the criteria used in classifying points had to rely on simple geometric
characteristics of a point relative to its neighbourhood. Methods for ALS data processing could be
much improved if ALS systems measure, in addition to the range, further physical observables which
can be used for vegetation classification. New ALS systems, which record the full echo-waveform,
may provide crucial information for the classification of vegetation points. In this paper we show that
the additional features derived from the full-waveform data - the amplitude, the pulse width and the
number of pulses - can be used to discriminate between vegetation and non-vegetation points without
using geometry information. Thus, a truly three dimensional representations of the classified ALS
points can be obtained. The classification algorithm is based on a decision tree technique. The
applicability of this method is demonstrated on data collected by the RIEGL LMS-Q560 sensor over
the Schönbrunn area of Vienna. The performance of the classification algorithm was checked
manually on 500 points randomly distributed and on several test zones selected over the study area.
We found an overall accuracy of 88.6% with a kappa coefficient of 0.8.

http://aleph.ub.tuwien.ac.at/F?base=tuw01&func=find-c&ccl_term=AC06586650

http://publik.tuwien.ac.at/files/PubDat_120277.pdf