You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

The 3rd ISPRS Workshop on Dynamic and Multi-Dimensional GIS & the 10th Annual Conference of CPGIS on Geoinformatics
Chen, Jun

ISPRS, Vol.34, Part 2W2, “Dynamic and Multi-Dimensional GIS”, Bangkok, May 23-25, 2001
uses on sustainable
3eijing to investigate
3 are involved in the
>lant Nutrition, Plant
rounding area of the
lacro scale. For the
, and the method of
sen for land use
a spatial information
rom satellite images
coordinate systems
dI points
>f land use map
For some maps in large scale it is very difficult or impossible to
get coordinates or information about projection, if such
information is available there are still problems to face. In Fig.1,
this becomes obvious, even so it does not show only large scale
scale data. The data do not match.
Furthermore, for a soil and land use map which was available for
the township in a scale of 1:10,000, the information about
coordinates and projection was not accessible. Therefore, the
acquisition of highly accurate x,y-coordinates for the
georeferencing of the IKONOS2 image and of all other available
spatial information is essential to ensure further analysis and
enable the evaluation of the used data.
For the collection of the high accurate x,y-coordinates, a DGPS-
system was set up for field work without relying on base stations
or any other data source in China. In Fig.2 the components of
the system are shown. For the GPS receiver, a GARMIN GPS
III+ was chosen due to its small size and its high compatibility for
DGPS receivers. Accuracy of this GPS is around 10 m since
May 2000. The DGPS receiver itself is a Fugro Omnistar
3000LM which operates with a 24 h satellite based DGPS
system (www.omnistar.com). For the data logger, a small
handheld computer, HP Jornada, with a 24 MB memory card
was used. The Garmin GPS III+ itself can store quite a number
of waypoints, too. Power supply for the GARMIN GPS III+ are
AA batteries. For the OMNISTAR 3000LM DGPS, a motor cycle
battery was used.
Fig.2: Components of the used DGPS system
The advantages of this system are its small size, low weight,
independence, and high accuracy of around 2 m. The
disadvantage is that the system is not able to store elevation
data which have to be written down separately. The accuracy of
the DGPS elevation measurements in the township of
Dongbeiwang is around 5 m. Fig.3 shows the components of the
DGPS system, too. The height of the GARMIN GPS III+ (1. in
Fig.3) is about 12 cm not including the 5 cm antenna of the GPS.
Units 1. to 4. in Fig.3 are necessary for the DGPS survey. Total
weight is around 4 kg. The battery charger is for the power unit
(4. in Fig.3) of the DGPS receiver (2. in Fig.3). The battery for
the DGPS receiver lasts for 30-40 working hours. The notebook
is only necessary to run special software for the first activation
of the DGPS license. For the processing of the DGPS data and
the creation of the polygons Fugawi3 (www.fugawi.com) and
Arclnfo7.2.1 (www.esri.com) are used. Finally, the photo in Fig.4
shows the DGPS system in the field during the survey in
Dongbeiwang in late summer 2000. Further data collection is
done in Spring 2001 for Dongbeiwang and selected areas of
Beijing Province.
2. DGPS receiver
3. DGPS antenna
4. Power source
for DGPS
Fig.3: Photo of the components of the DGPS system
Fig.4: DGPS system in “action”
The first step of the IKONOS2 image processing, which was
ordered in GEO50 quality due to costs, is the orthorectification
and georeferencing. The GEO50 accuracy is 50 m (RMSE 23,6
m) (www.spaceimaging.com). Using the ground control points
(GCOs) of the DGPS mapping, the IKONOS2 image was
resampled with a 1 m resolution for the panchromatic file and a
4 m resolution for the multispectral files. The accuracy is around
2 m. This accuracy, deriving from the DGPS mapping, is
comparable to IKONOS2 High Precision Plus data which are not
available for China. The Software for the processing of the
images are IDRISI32 (www.clarklabs.org) and Arclnfo7.2.1 for
Windows2000 on a Toshiba Tecra 8000. ArcView3.2 is mainly
used for display the results.
In Fig.5, the results of the first image processing are displayed.
The displayed area in Fig.5 covers the surroundings of the
experiment field in Dongbeiwang, which covers around 15 ha.
Some of the research plots of the field experiment are also
shown in Fig.5 as well as the results of the DGPS mapping (red
points and lines) and the georeferenced soil and land use map
1:10,000 of this township (yellow lines). The GCOs of the DGPS
mapping were also used for the georeferencing process of this
land use map