te and 
survey 
Ty fast 
ng the 
moved 
mplete 
points 
single 
) avoid 
s been 
| to the 
an be 
irectly 
domly 
cts the 
sity of 
step of 
ted by 
oosen, 
cloud 
yr each 
nd the 
bitrary 
by the 
al and 
to the 
actions 
of the 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
object, starting from a set of point clouds. This set of actions 
can be divided into 2 different stages: 
e the pre-treatment (or preliminary treatment) of the 
laser data; 
e the solid modelling of the point cloud. 
As “preliminary treatment” we mean all the operations that 
are directly carried out on the point cloud, such as, for 
example, the data filtering (noise reduction), the point clouds 
registration and georeferencing operations. The result of 
these procedures is a complex "noise free” point cloud 
(without outliers, gross and systematic errors) and this is the 
correct starting point for the second stage of the treatment of 
laser data: the 3D modelling. 
The second part of the laser scanner data management, the 
3D modelling, is a set of operations that, starting from any 
point cloud, allows a surface model of the object to be 
formed. 
While there is a huge range of different products on the 
market to carry out solid modelling, just few software for a 
correct. preliminary treatment of terrestrial laser scanner data 
can be found. 
3. THE LSR 2004 SOFTWARE 
3.1 The LSR 2004 
LSR 2004 (Laser Scanner Registration) is a software package 
for the preliminary laser scanner data treatment that has been 
developed by the authors. This software was written 
completely in Visual Basic 6.0 language and manage also 
some products directly obtainable form point clouds without 
any modelling phase.(e.g. Solid Image, Orthophoto and Truc 
Orthophoto production). 
  
  
  
  
  
  
  
  
  
  
Figure 1. The LSR 2004 software package 
LSR 2004 allows: 
* point clouds and high resolution digital images to 
be imported and exported; 
* acquired point clouds to be filtered in order to 
remove outliers and gross errors; 
* position of the high reflectivity points (markers) 
arranged on the acquired object to be determined 
automatically; 
* adjacent scans to be registered and/or 
georeferenced automatically using reference 
markers; 
e laser triangulation to be carried out in the case of a 
set of scans; 
e digital Images to be calibrated; 
e solid Image and its applications (3D colour model 
and artificial stereoscopic pair) to be created and 
managed; 
e direct measurements to be carried out on the Solid 
Image. 
Some of these procedures are here described. 
3.2 Search for the high reflectivity points 
Modern laser scanner devices record, for each acquired point, 
the direction of the laser beam (horizontal and vertical angles 
are recorded), the measured distance and the reflectivity 
values (e.g. the energy reflected by the measured point). This 
set of information make possible to calculate the 3D 
coordinates of each point using well known geometric 
equations. 
Reflectivity information is still not frequently used in 
commercial software, but it should not be ignored when 
dealing with research or specific applications. The reflectivity 
value is in fact connected to the type of material that makes 
up the object, an aspect that can be of fundamental 
importance in the analysis and development of automatic 
algorithms or in a first calssification of the materials. 
One of the possibilities offered by knowing this set of 
information is the opportunity of automatically register (or 
georefer) two adjacent point clouds. To do this, it is sufficient 
to arrange some high reflectivity stickers (markers) on the 
object during the scan. 
When the laser beam strikes the markers, the recorded 
reflectivity value is very high and is usually much higher than 
the one recorded on natural points (for example, rock, wood 
etc). If this simple property is used, it is possible to 
automatically identify the position of the markers inside the 
3D model acquired with the laser scanner. 
The markers should however be suitably sized and arranged 
on the object if a correct determination of their position is to 
be obtained. Their position and size have to be planned 
considering the laser-object distance, the used angular 
resolution and the mean inclination of the measuring 
directions. 
  
Figure 2. The markers 
A specific search algorithm has been implemented inside the 
LSR 2004 software. This algorithm runs in two subsequent 
steps. 
First of all, all the points with a reflectivity value higher than 
an imposed minimum are identified inside the 3D model. 
This value can be manually established by the operator or 
automatically obtained by the computer. In this case, the 
minimum value imposed for the point search is determined 
by evaluating the reflectivity histograms of all the 3D model 
points. 
The second step of the proposed algorithm may possible 
clustering the identified points in order to group all the points 
that describes the same reflecting marker. In fact a marker 
positioned on the object can be struck more than once by the 
laser beam. For this reason it is necessary to group those that 
refer to the same marker together. The position of each 
marker is calculated through a weighted mean (in function of