VE
EEE
Pseudo-Grid Based Building Extraction Using Airborne LIDAR Data
Woosug Cho **, Yoon-Seok Jwa®, Hwi-Jeong Chang ®, Sung-Hun Lee"
* Associate Prof. Dept. of Civil Eng., Inha University, Inchon 402-751, KOREA - wcho&zinha.ac.kr
? Graduate student, Dept. of Geoinformatic Eng., Inha University, Inchon, KOREA -
oon1314@1ycos.co.kr
KEY WORDS: LIDAR, Building Extraction, Building Detection, Airborne Laser Scanning, Pseudo-Grid
ABSTRACT:
This paper proposed a practical method for building detection and extraction using airborne laser scanning data. The proposed
method consists mainly of two processes: low and high level processes. The major distinction from the previous approaches is that
we introduce a concept of pseudo-grid (or binning) into raw laser scanning data to avoid the loss of information and accuracy due to
interpolation as well as to define the adjacency of neighboring laser point data and to speed up the processing time. The approach
begins with pseudo-grid generation, noise removal, segmentation, grouping for building detection, linearization and simplification of
building boundary, and building extraction in 3D vector format. To achieve the efficient processing, each step changes the domain of
input data such as point and pseudo-grid accordingly. The experimental results show that the proposed method is promising.
1. INTRODUCTION
In recent years, accurate 3D data in urban areas is in great
demand for many applications such as urban planning, mobile
communication, 3D city modeling and virtual reality. Usually
urban areas are dynamically changing due to construction and
extension of urban features, especially buildings. Detection and
reconstruction of buildings are of highest interest in the
geospatial community. Since manual digitizing is time
consuming and very costly, a fast and automated method for
detecting and extracting buildings is required by many users of
geographic information system (Palmer, 2001).
Airborne laser scanning is a relatively new and promising
technology for obtaining Digital Surface Models (DSMs) with
high density and high positional accuracy of the earth surface.
The development of airborne laser scanning started in the 1970s
(Schenk, 1999). Airborne laser scanning system comprised of
laser scanner, GPS receiver and IMU computes the range to the
target point by emitting a laser pulse and measuring the round-
trip time. Contrary to the passive sensor such as optical sensor,
the laser scanner is an active sensor so that it works day and
night, and is less affected by the shadow and weather condition
(Baltsavias, 1999).
A number of research works have been performed on building
detection and reconstruction from airborne laser scanning data
in automated fashion. Wang (1998) used the shape information
to separate buildings from all other objects based an assumption
that most buildings have simple and regular shapes and other
objects do not have. The shape information is obtained from
edges detected on the elevation image in regular grid that is
converted from laser data. In Maas and Vosselman(1999), the
authors presented two techniques for determining parameters of
gable roof type building models from laser altimetry data. Both
techniques work on the original laser scanner data points
without the requirement of an interpolation to a regular grid.
Wang and Schenk (2000) reported an approach that takes high
quality terrain surface data generated by airborne laser scanning
data as input and goes through edge detection, edge
classification, building points extraction, TIN model generation,
and building reconstruction to extract and reconstruct buildings
and building related information. For building detection, it
detects edges from the surface data and classifies edges to
distinguish building edges from other edges based on their
geometry and shapes including orthogonality, parallelism,
circularity and symmetry. Morgan and Tempfli (2000)
developed a procedure starting by resampling elevation as
obtained by laser scanning into regular grid. The core part of
building detection is based on a morphological filter for
distinguishing between terrain and non-terrain segments. The
non-terrain segments are classified into building or vegetation.
In Morgan and Habib(2001), the authors generated a 3D TIN
structure using the irregularly distributed laser data for building
detection and extraction. The 3D TIN is generated to serve the
detection of the building facades including the vertical walls.
Elaksher and Bethel (2002) utilized the geometric properties of
urban building for the reconstruction of the building wire-
frames from the LiDAR data. The approach started by finding
the candidate building points that are used to populate a plane
parameter space and followed by filling the plane parameter
space, extracting roof regions and refining the plane parameters.
Finally, the region boundaries are extracted and used to form
the building wire-frames.
In this paper, we propose a practical method for building
detection and extraction in urban areas. The proposed method
consists mainly of two processes: low and high level processes.
The major distinction from the previous approaches is that we
introduce a concept of pseudo-grid (or binning) into raw laser
scanning data to avoid the loss of information and accuracy due
to interpolation as well as to define the adjacency of
neighboring laser point data and to speed up the processing time.
The practical approach proposed in the paper begins with
pseudo-grid creation, noise removal, segmentation, grouping for
building detection, linearization and simplification of building
boundary, and building extraction in 3D vector format. To
achieve the efficient processing, each step listed above changes
the domain of input data such as point and pseudo-grid
accordingly. Figure 1 illustrates the schematic diagram of the
practical approach for building detection and extraction
proposed in this paper.
Internat
low- 1
proce
high- |
proci
Figure
buildin;
The prc
and hi;
pseudo-
high le
boundai
of inpui
in order
Figure
approac
high- le
proces
low- le:
proce:
i
Figure
for buil
