Figure 1-2 3D ortho Product generation algorithm flow 
It’s difficulty to measure the Ground Control Point (GCP) by 
GPS or other method in the West Region of China where the 
1:50,000 scaled terrain maps are absent, because of the 
mountainous terrain, rugged topography and abominable 
climate of those area. So it’s an important method to make the 
topographic map using the high-resolution remotely sensing 
imagery with high precision orientation parameter to extract 
DEM under the condition of lacking GCP, then the DEMs were 
used to make the ortho image. 
Up to till now, main emphasis has mainly given to the analysis 
of geolocation error of remote sensing imagery. Zhang 
Jixian(2000) researched the displacement of SPOT images, 
M.S.Hashemiana(2004) evaluated the accuracy of DEM gener 
ated form SPOT5 HRS imagies, Philip Cheng() described how 
to correct QuickBird data using different geometric correction 
model and the method of automatic DEM extraction using 
QuickBird stereo data. Wang Yali(2005) analyzed the SPOT5 
geometric error applying the computer simulation technique, 
and A. M. Marangoz (2005) researched the geometric evalua 
tion, automated DEM and orthoimage generation from along- 
track stereo ASTER Images. 
The aims of this paper is to assess the geolocation accuracy of 
ASTER L3A image. If the accuracy satisfies the requirement 
of China’s 1:50,000 map, then the ASTER image can be used 
in the Mapping Project in the West Region of China. 2 
2. THE ACCURACY ASSESSMENT OF ASTER L3A IN 
BEIJING AREA 
2.1 the terrain of Beijing area 
Beijing lies in the north of the North China Plain, at 39° 56' 
N and 116° 20' E. The terrain of the Beijing area slopes from 
the northwest to the southeast. Mountains snake round the 
city's north, west and northeast, while the southeast part of the 
city is a plain that slopes gently toward the coast of the Bohai 
Sea. The Yongding, Chaobai and Juma rivers and the north 
section of the Grand Cannel crisscross the area under Beijing's 
jurisdiction. Most of the rivers originate from mountainous 
areas in the northwest, cut through mountains and zigzag 
through the plain in the southeast before emptying into Bohai 
Sea. It covers 16,807.8 square kilometers. Mountainous areas 
occupy 10,417.5 square kilometers, accounting for 62% of the 
city's landmass. The rest, 6390.3 square kilometers or 38% of 
the total, are flatland. 
Figure 2-1 the terrain of testfield in Beijing area and the 
distribution of check points 
2.2 assessment approach 
The geometric accuracy validation approach of ASTER Level- 
3A image is that choosing character points form the image first, 
then measuring the ground coordinates of the character points 
by Global Positioning System (GPS), and the geolocation 
accuracy was got by comparing the coordinates measured by 
GPS and the image coordinates. 
2.2.1 choose the character points in the image 
The swath of ASTER image is 60 km, so we decided to choose 
40 character points distributed well. Same with many satellite 
datasets: road crossings and the end of bridges can easily be 
selected as character points in ASTER scenes. In mountainous 
area, it’s difficulty to choose the character points because there 
are few roads and the roads always are not wide enough ( the 
ground resolution of ASTER image is 15m) to recognize, so 
other points such as turning points of a road was chosen as 
character points. The distribution of character points is shown 
in figure 2-1. 
2.2.2 measure the coordinates of the character points 
The ground coordinates of the 40 character points were meas 
ured by GPS, and planimetric accuracy of these points can 
reach to 1 m. 
2.2.3 compare the coordinates difference 
268