International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
2.3 Modeling of Points Clouds 
An explicit useful spatial dataset is achieved by jointing all 
point clouds of multiple scans of a targeting object The vectors 
can be projected onto a plan surface or transformed into a map, 
or even to make a 3D perspective views. 
2.4 Coordinate Transformation of Points Clouds 
After joining all datasets into a single point cloud model with a 
common coordinate system, a further step is to transform the 
local coordinate systems into a geodetic system by using 
ground control points measured by ground surveying, in this 
study using GPS and Total Stations. 
  
| 
   
    
    
X x-X, 
Y > Rens : = V 
(X, Y. Z:) s = t 
V X 
A Scanner Coordinate System * 
Objed Coordinate System — X 
  
  
  
Figure 2.2. A sketch showing the coordinate transformations 
of multi-scanning point clouds 
3. RESULTS OF THE CHANGE DETECTION 
In this study, ground-based 3D laser scanning is applied to two 
cases. One is a river-side landslide in Chudong of Hsin-chu 
County which is a typical and common one in Taiwan (Hsiao et 
al., 2003a & 2003b). The other is the toe area of the large-scaled 
dip-slope landslide of Jiu-fen-er mountain in central Taiwan. 
3.1 The River-side Landslide Case 
The Chudong landslide is located in a small valley with a very 
limited open space for taking a good view of the whole area 
(Figure 3.1). Six scanning stations with scanner-to-target 
distances less than 100 m were applied. 
  
    
Figure 3.1. The area covered by ground-based lidar with 6 
stations. 
The extent of a scanning station and the point spacing as well as 
the duration required for a scan are displayed on the PDA 
controlled panel of the scanner ILRIS-3D. Six datasets 
generated by six scanning stations are subsequently merged into 
one single dataset. And, then, a coordinate transformation was 
conducted by using six control points distributed in the area and 
measured by GPS (Figure 3.1). Thus, a useful topographic 
model is established. Figure 3.2 shows the application of a 2D 
plane map and a 3D perspective map. 
For understanding the topographic change of the landslide area, 
an existing topographic map in a scale of 1:5000 was digitized 
to obtain the topography before the sliding event. Though no 
other more detailed topographic information could be accessed 
for the study area, the most common topographic information of 
the 5K photomaps are useful for a quick understanding of the 
618 
landform evolution of the study area. As shown in Figure 3.3, 
the topographic information from 5K topographic maps and that 
from laser scan are compared to obtain the altitude variation. A 
volume change with obvious spatial variations indicates a 
significant result can be achieved by this approach. The 
estimated change of earth volume is 11,691 m? with a landfill of 
4,899 m? and a cut of 6,792 m°. 
Additional information, a merit in contrast to the traditional 
topographic survey, is the portrait of detailed profiles of the 
landslide using the densely distributed data points, as shown in 
Figure 3.4. 
  
  
  
  
  
(a) 2D plane map (b) 3D perspective map 
Figure 3.2. 2D plane and 3D perspective maps after 
modeling of points cloud data 
  
  
  
  
  
  
    
  
  
  
  
  
  
  
  
  
  
(a) 5 profile map 
  
  
  
(b) A profile altitude |(c) B profile altitude 
N 
VEI ; L4 mms er ; 
ru HE S SO ; iS. 
  
  
  
  
  
  
  
  
  
  
  
  
  
(d) C profile altitude. | (e) D profile altitude | (f) E profile altitude 
  
Figure 3.4. Profiles of the Chudong landslide 
32 The Jiu-fen-er Slide Area 
The most common digital terrain information in Taiwan is the 
DTM with a 40 m grid. Another most common topographic 
information is the 5K photomaps. Another datasets for this area 
is the point clouds taken by airborne lidar (Shih et al, 2002). 
Stereo-pairs of aerial photographs of the area are used to 
generate | m contour interval topography. Figures 3.5 and 3.6 
are the study area and its DEM modeling. Figure 3.7 is a 3D 
perspective view generated by from SPOT image and 40m 
DTM. Figure 3.8 shows the topography before and after the 
Ji-Ji Earthquake. A dramatic change of topography can be 
perceived. 
Airborne lidars of Optech ALTM 2030 and LH ALS40 were 
introduced to Taiwan by Agriculture Council for taking test 
Inter 
  
datas 
2002 
were 
lidar 
In su 
been 
of th 
show 
post- 
taken 
Figut 
airbo