The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part BI. Beijing 2008
278
2.2. Test region and test method
There are many mountain massif landslide areas in Jingyang,
because of natural disaster or long-term change of
circumstance. Some landslides are more than several
kilometers long and ten meters height. Figure 1 and Figure 2
are the landslides of this region; Figure 2 shows the ground
fissure because of the landslide. If the landslide is near the
residential area, more people will be threatened.
Figure 1: The landslide Image
Figure2: Ground fissure because of landslides
In this test, a typical area with 50 meters height and 100 meters
long was chosen. It was scanned twice with a fixed station. In
the end, it acquired the 3D point cloud and RGB images of the
test area. In order to simulate the possibility of the landslide
monitoring by this system, a lab assistant is considered as a
target on the landslide area. This tested person moved about 1
meter along landslide direction before the second scan. It is a
simulated landslide by inspecting the position of tested person.
The Basic parameter of the data collection is listed in the table
2.The Figure 3 shows the image of data collection region and
the position of the tested person. The yellow circles show the
position of the tested person.
Scan distance
100 meter
Horizontal angle
50
Vertical angle
50
Number of shots
5453630
Number of scan lines
1690
Number of points per line
3227
Pulse mode
First Pulse
Scan start position
Lower Left
Scan orientation
Horizontal
X spot spacing(counts/mm)
8/35.2
X spot spacing(counts/mm)
8/35.2
Table 2: The data collection of landslide area in Jingyang
2.3. Purpose of test
There are two purposes in this test. Firstly, it simulates the test
area of landslide by the position change of the tested people
and gains quantificational result. Secondly, it should filter
other objects outside the tested people such as the noise and the
vegetation, and then, we can contrast the results of two scans
and get the accuracy of the landslide monitoring.
Figure 3: The image of data collection region and the position
of the tested person
3. DATA POST PROCESSING AND RESULT
3.1The results of the data post processing
The original datum should be converted into pif format, this
format data can be read by Polyworks. The datum is processed
by Poly works and Terrascan. Terrascan is used to the point
cloud filtering here. Polyworks is used to compare to the twi-
scanned point cloud and gets the result of landslide.
The datum is processed by some module of Polyworks, Data
processing workflow is as follows (Innov Metric Software,
2003):
1) Firstly, the collected points should be processed to grid
by IMalign module of Polyworks. They didn’t need align
because of only one fixed station to scan. If the area is
large, it should be set several stations from different angle
and align all 3D data into unique coordinate system.
2) Export pqk format model.
3) The point cloud should be compared in Inspect module.
The first scanned point cloud are the reference one. The
second scanned point cloud are the compared one, which
will compare with the first scanned point cloud.
The test result is showed in Figure 4, Figure 5 and Figure 6.
Figure 4 shows that the result of overlapping of scanned point
cloud. Figure 5 shows the result of visualizing comparison
results as colors mapped on the point cloud. Because the