591
TLS DEFORMATION MEASUREMENT USING LS3D SURFACE AND CURVE
MATCHING
O. Monserrat, M. Crosetto, B. Pucci
Institute of Geomatics, Castelldefels, Barcelona, Spain, - (oriol.monserrat, michele.crosetto, barbara.pucci)@ideg.es
KEY WORDS: Terrestrial Laser Scanner, Point Cloud Matching, Deformation measurement, Precision, Curve Matching
ABSTRACT:
During last few years the use of Terrestrial Laser Scanner has increased notably in different application fields, like architecture,
geology and geodesy. The paper focuses in the use of TLS data for deformation measurement and monitoring purposes which
concerns both engineering geology and geodesy. The paper presents a new approach for deformation measurement which fully takes
advantage of the TLS data characteristics. The procedure is based on the point cloud matching algorithm Least Square 3D Surface
Matching proposed by Gruen and Akca (ISPRS Journal, 2005, 59, 151-174). The proposed approach takes advantage of both surface
and curve matching to improve the co-registration quality and it exploits the high density of TLS point clouds, which counter
balances the relatively poor precision of the single TLS points. The work describes the proposed deformation measurement
procedure, and discusses in particular some results of the research made at the Institute of Geomatics: the preliminary results
obtained with curve matching.
1. INTRODUCTION
In the last years, the Terrestrial Laser Scanner (TLS) has been
an increasing interest as a method for deformation measuring in
different fields such as engineering geodesy (Alba et al., 2006
and Schneider, 2006) and several applications of geology and
geotechnics like landslide monitoring (Bitelli et al., 2004) and
rock falls (Alba et al., 2005). At the same time the availability
of different typologies of TLS instruments has increased. A
classification of available TLS can be found in Frohlich and
Mettenleiter (2004). This increasing interest is due to the key
advantages of the TLS with respect to classical techniques, e.g.,
the high sampling density, the portability and the automatic 3D
point measurement. The first one is a key point to counter
balance the relatively poor quality of the single TLS points and,
as a consequence, to measure deformations.
The paper presents a new methodology for land deformation
measurement from TLS data. This procedure takes advantage of
the high redundancy of the point clouds for estimating the
deformation parameters. The core of the procedure is the point
cloud co-registration based on the least squares 3D surface and
curve matching proposed by Gruen and Akca in 2005.
The work is organized in three main parts. Firstly is described
the proposed deformation measurement procedure which
includes three main steps, the acquisition of the data, the global
processing of the entire scene and the estimation of the local
deformation. In the second part the first results obtained with
the proposed procedure are presented, and finally the
preliminary results obtained with curve matching are discussed.
Some conclusions follow.
2. PROPOSED APPROACH
The proposed approach takes full advantage of the geometric
information contained in the point clouds acquired during
different epochs, in order to maximise the sensitivity to terrain
changes. The core of the procedure is the least squares 3D
surface matching described in Akca and Gruen (2005) and
Gruen and Akca (2005), which has been implemented by the
Institute of Geomatics. The procedure is flexible and can be
used in a wide range of applications. Figure 1 shows the data
flow of the whole procedure, whose main steps are described
below.
TLS acquisitions
♦ | Remove movement areas
Figure 1: Data flow of the proposed procedure. Its key points
are highlighted in bold.
1. Acquisitions of the TLS data: Let’s assume to have a
deformation area to be monitored, which is surrounded by a
stable area. The data acquisition involves at least two steps: