Accuracy without ground control requires accurate knowledge of orbital ephemeris, satellite attitude (pointing direction), 
and interlock angles between the telescope bore-sight and star trackers. On-board GPS receivers and ground processing 
determine orbital ephemeris. Star trackers determine satellite attitude. Combining ephemeris, attitude, and interlock 
information enables ground processing to triangulate on ground features and determine their position. The requirements in 
Table 2 on 1, 3, and 5 models are three specification points on a curve of asymptotically improving accuracy as the number 
of models is increased. 
Number of Stereo Models Horizontal A , CE90 Vertical ,LE90 
1 25.0m 22.0m 
  
3 12.2m not 
5 not i 10.0m 
Table 2 Metric Accuracy Specification without Ground Control 
3.3.2 Stereo Accuracy with Ground Control 
Use of Ground Control Points (GCP) improves accuracy, largely by eliminating bias errors. The requirements for stereo 
model accuracy with ground control are shown in Table 3. 
  
Horizontal Accuracy 2m CE90 
  
  
  
  
Vertical Accuracy 3m CE90 
  
  
Table 3 Metric Accuracy Specification with Ground Control Points 
3.3.3 Operational Metric Accuracy Verification 
Space imaging uses ground test ranges to verify the operational metric accuracy of the system. Test ranges have been 
surveyed in Phoenix, San Diego, and Western Australia. These geographic regions provide geographic diversity, differing 
terrain conditions and high probability of cloud free imaging. The Phoenix test range consists of 150 photo identifiable 
ground control points, GCP’s. The points are fairly uniformly distributed over a 30km East-West by 20km North-South 
area. This range permits a mapping block of 3 side-by-side image strips. The San Diego test range consists of 140 GCP 
over a 22 by 22km area and offers more vertical relief, hence a more rigorous test of terrain extraction and 
orthorectification. The Southern Hemisphere Test Range consists of 200 GCP’s over a 100 by 100km area in Western 
Australia. This area is relatively. Located in the Southern Hemisphere, this range is an important test that accuracy is 
independent of position along the orbital path. Being in the Eastern Hemisphere also tests that accuracy is independent of 
longitude. The GCP layout for this test range provides for measuring a larger mapping block, as well as independent 
analysis of pixel field angle mapping and along track low frequency oscillation. Table 4 shows the accuracy for Carterra 
  
  
  
  
  
  
  
Carterra CE90 RMS NMAS '!' 
Product 
Geo ^ 50.0 m 23.3 m 1:100,000 
Reference 25.4 m 11:8 m 1:50,000 
M ap 12.2 m 5.7 m 1:24,000 
Pro 10.2 m 4.8 m 1:12,000 
Precision 4.] m 1.9 m 1:4,800 
Precision Plus 2.0m 0.9 m 1:2-400 
  
Table 4 CARTERRA products and associated metric accuracy e 
Notes: 
(1) Meets National Map Accuracy Standard for scale shown. Pixel size is 1m. 
(2) Geo products are not orthorectified; accuracy values do not include effect of terrain displacement. 
(3) Not all products in the table have been released as of the date of publication. See section 4.1 
products, both released and planned. 
  
134 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B1. Amsterdam 2000. 
Pro: 
IKC 
1), 
2) 
3) | 
4) | 
Pro 
gra) 
is th 
Futi 
colc 
elev 
DE! 
4.1 
Fori 
Proj 
Dati 
Mec 
42 
On. 
IKC 
deli 
orth 
disp 
disp 
disp 
CES 
4,3 
Cart 
inch 
site 
Ref: 
the : 
orth 
Cote 
spec 
4.4 
Cart 
utili 
app! 
8€0; 
prod 
accu 
the i 
orth