Full text: Proceedings; XXI International Congress for Photogrammetry and Remote Sensing (Part B1-1)

427 
COMPARISON OF TREE EXTRACTION FROM INTENSITY DROP AND FROM 
MULTIPLE RETURNS IN ALS DATA 
C.Ormeci a , S.Cesur b 
a ITU, Civil Engineering Faculty, 80626 Maslak Istanbul, Turkey - ormeci@itu.edu.tr 
b ITU, Informatics Institute, 80626 Maslak Istanbul, Turkey - cesurs@itu.edu.tr 
KEY WORDS: LIDAR data, Tree extraction, Airborne Laser Scanner, Three dimensional, Intensity Drop, Multiple Returns 
ABSTRACT: 
LIDAR is an application widely open as an area of study for modelling detailed topographic maps. It is generally used in airborne 
applications. In this study, the aim is tree extraction from a data having the 3D coordinates of first and last returns, and intensity data. 
The data used is from the city center of Stuttgart/Germany and includes large and irregularly shaped buildings and road with a bridge 
and tunnel. Two methods were compared. In the first method, the algorithm searched for neighboring points with low intensities. The 
points with lower intensity than the treshold represented trees, as trees cause absorption. In the secon method, the algorithm searched 
for the points that had height difference between the first and last returns. These point clouds were effective on tree extraction, as 
trees cause multiple returns. The density of trees in the first method was poorer than in the second, while in the second method, 
balconies and comers of the buildings were seen as trees. As a result, the second method is better on tree extraction, however an 
additional edge-detecting filter is needed. 
1. INTRODUCTION 
LIDAR is acronym for Light Detection and Ranging and is used 
to gain high resolution topograghic maps more accurate than 
traditional methods. It has the same working principle as radar, 
except using laser light instead of radio waves. A typical Lidar 
consists of transmitter (one laser or more), transmitter optics, 
receiver optics, detector and electronic system. Lidars may be 
classified according to process (range finders, DIAL, Doppler 
Lidar), platform (terrestrial, airborne, spacebome), wavelength 
(infrared, visible, ultraviolet), scattering type (Backscattering 
Lidar, Rayleigh Lidar, Raman Lidar, Fluorescence Lidar) 
(Weitkamp, 2005). Airborne Laser System is one type which 
the Earth is scanned from an airplane or a helicopter and the 
surface elevation is estimated by return time. A small telescope 
(detector) and a mirror (receiver optics) is used to collect the 
light back. The position of the plane is determined by a 
differential Global Positioning System (dGPS). The orientation 
of the plane is measured by an Inertial Navigation System (INS) 
(Figure 1). 
z 
OB' 
Figure 1. ALS and its components [2] 
All these information is collected to have a “geolocated laser 
return” (Harding, 2000). The ground that is to be scanned can 
be rocky mountains or urban city or densely forested area. The 
produced data consists of dense point clouds and until now, 
several filters have been developed to interprete bare Earth, 
buildings and vegetations; morphological, or segment-based, or 
slope-based, or iterative and or point to point (Sithole, 2004). 
In this paper, two methods were used for tree extraction. The 
raw data used is test site 2 in the ISPRS Working Group III/3 
and belongs to Stuttgart city center (Figure 2). 
Figure 2. Stuttgart city center. (Google Earth Application Image 
Copyright 2007 GeoContent All Rights Reserved )
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.