International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
In this study, non-horizontal terrain areas are also selected and 
tested, e.g. two gable roofs shown in Fig. 4. The coordinate 
differences of non-gable corners are not very obvious. From 
Point number 7901192 and 4861413 in Table 1, it shows that 
max absolute is 0.31m, 0.31m, and 0.60m in X, Y, and Z 
coordinate components respectively. Two distinct gable corners 
are selected and compared from orthoimage patch 7901192, 
their coordinate difference in X,Y are only (0.23m, 0.07m) 
and (0.42m, -0.04m), but the Z difference reached -1.14m 
and -1.71m. A gamble corner in orthoimage number 
4861413 is measured and compared. It's difference is 0.26m, 
0.01m, and -1.23m in X,Y, and Z direction. Obviously, the 
height difference is large. But the X,Y coordinate 
differences are not so large from these results. 
  
B. Orthoimage patch of terrain area 4861413 
Figure 4 The orthoimage patch from non-horizontal terrain area 
During the process, some difficulties are found in this proposed 
methodology. That means how to (1).find a time-invariant flat 
terrain area (2).obtain well-contrast orthoimage patches with 
suitable size and texture. Actually it is very difficult to find lots 
of terrain areas that meet the (1) and (2) requirements. But the 
reachable geometrical accuracies motivate us to do more studies. 
For example, it is necessary to develop a mechanism to evaluate 
whether an orthoimage patch is suitable for criterion (2) and 
self-verify the output results. Namely, the criteria for self- 
examination of geometric and radiometric quality should be 
developed in order to ensure what the operator choose is good 
enough for controls. 
Additionally, the evaluations of radiometric quality and the 
accuracy assessment of exterior orientation by using these 
orthoimage patches as controls should be done in the future. 
Meanwhile, tests by matching new aerial images with 
orthoimage patches for orientation parameters no matter what 
interactively or automatically should be done. 
Some improvements could be also conducted for generating 
orthoimage patches much more efficient and more accurate, for 
example, the selection of suitable terrain area and the provision 
of height approximation. In this study, only the approximation 
of height information is used for the orthorectification. No 
precise height information must be measured by the operator. In 
the future test, digital surface model (DSM) could be included 
880 
for provision of approximation height information. Additionally, 
the radiometric model in Data Snooping can be improved more 
accurate. 
Finally, horizontal plane is basic assumption, this is too limited. 
The test for non-horizantal plan or the refinement of the result 
from generated orthoimage patches can be done in the future. 
S. CONCLUSIONS 
A methodology is proposed to generate the orthorectified image 
patches as controls for aerotriangulation in this paper. Even 
further investigation and study should be conducted in order to 
get more verification, the preliminary tests show that the RMSe 
can reach 0.15m, 0.19m, and 0.29m in X,Y, and Z coordinate 
components. The good geometric accuracy proves the 
feasibility of the proposed method for the controls in aerial 
triangulation. 
REFERNENCES 
Baarda, W., 1968, A Testing Procedure for Use in Geodetic 
Networks, Netherlands Geodetic Commission, Delft, the 
Netherlands. 
Collins, R.T.,1996. A Space-Sweep Approach to True Multi- 
Image Matching. In: [EEE Computer Vision and Pattern 
Recognition (CVPR96), San Francisco, CA, June 18-20, 1996, 
pp. 358-363. 
Hóhle, J., 1999, Automatic Orientation of Aerial Images on 
Database Information. In: OEEPE Official Publication No 36, 
pp. 71-192, ISSN0257-0505 
Hóhle, J., 2001. Automated Georeferencing of Aerial Images. 
In: Proceedings of ScanGIS'2001, Às, 
Jin, J, Z. Wang, and Q. Peng, 2000. An Optimal Fusion 
Method of Mulei-spectral Images under Correlation Coefficient 
Criteria. In: Journal of Computer-Aided Design & Computer 
Graphics. 
Paszotta, Z., 2000. Method of Exterior Orientation of Aerial 
Images by Matching Orthoimages. ISBN 83-88343-79-3, 
WYDAWNICTWO  UNIWERSYTETU WARMINSKO- 
MAZURSKIEGO, Olsztyn. 
Shan, J., 1999. Automatic Exterior Image Orientation with 
Orthoimage and DTM. In: OEEPE Publication No 36, pp.151- 
167. 
Wolf, Paul R. and Bon A.Dewitt, 2000. Elements of 
Photogrammetry with Applications in GIS. 3rd Edition , 
McGRAW-HILL international edition, pp.383. 
Wolf, Paul R. and Charles D. Ghilani, 1997. Adjustment 
Computations: Statistics and Least Squares in Surveying and 
GIS. John Wiley & Sons, Inc., pp.402-406. 
Yang, L.-F., 2003. Automatic Exterior Image Orientation Using 
Existing Orthoimage and DTM. Master's Thesis, Dept. of 
Surveying Engineering, National Cheng-Kung University, 
Tainan, Taiwan, R.O.C. 
ACKNOWLEDGMENTS 
I am grateful to China Engineering Consultants, Inc. Taipei, 
Taiwan, that provides the test data set in this study. 
IN 
KEY 
ABS 
The 
focu: 
Ever 
meth 
is ne 
catec 
To fi 
befoi 
perfc 
cities 
An I 
phot: 
Posi 
by [ 
Proje 
univ 
aims 
and 
Qual 
is of 
and | 
exha 
the | 
Stutt 
acqu 
nece: 
attair 
neve; 
quali 
relial 
This 
whic 
paral 
imag 
the c 
and | 
diffe; 
andr 
To p 
diffe;