nbul 2004 
6 models 
ation less 
36/7, with 
rientation 
is parts of 
describes 
  
  
  
Figure 7 
w method 
1g for the 
ve Orientation 
  
i RMSE 
  
64 00064 
  
M] 04018 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
jf. digital 
ed on the 
xd among 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
Performances by digital drawing were mixed using 
measurement of reference points between adjacent models on 
MicroStation. Figure 8 shows elevation generated by editing 
performances of digital drawings in two dimensions related to 
front plane of an object using VirtuoZo system.. 
5. PRODUCTION OF ORTHO IMAGE 
Orthographic images mean images acting as plane figure as 
showing central projection objects on orthographic position. 
  
   
  
= xa 2 
ak 
Single photos per each Model GCP Acquisition 
  
Y 
Preprocessing = 
Image Histogram Matching 
    
   
  
Y 
Photogrammetry Processing 
Interior, Relative Orientation Epipolar geometry 
Resampling, Absolute orientation 
ES 
  
    
   
  
  
  
    
Y 
; Verification of Result / Postprocessing 6 
Filtering of DEM (Drape Image) Interactive Editing 
  
  
  
  
QUEM, UA ias 
  
  
  
Y 
| Final Results 
VEDI TAS CAR Y oe Hp A 
| Ortho Image Mosaic 
Ortho Image 
generation 
  
  
  
  
  
  
  
  
ARI 
Figure 9 Orthographic Image Generation Flow 
That is, orthographic images have only one difference 
that they express an object map with symbols and lines by 
images of an object so that they can be used like a map in 
directly measuring distances, angles, locations and 
dimensions. Like general aerial photography, close-range 
photogrammetric survey results are also influenced by 
central projection principles. Then, like performances of aerial 
photography in which objects are subject to displacement 
because of geography and landform, orthographic images 
also have some displacement according to landform. Thus, 
acquiring orthographic images and applying them in close 
photographic surveys will improve efficiency 
of surveys. Therefore, this study tries to generate orthographic 
images using orthographic rectification approach related to the 
main building of Korea University and then perform 3D 
modeling on the basis of the same coordinates system using 
acquired DEM and drawing performances on the front plane of 
an object. Figure 9 describes general process for creating 
orthographic images. 
5.1 Digital Images and Reference Points 
6 sheets of digital image data on the front plane of an object 
were used for creating orthographic images. Reference points 
were selected as specific points among vector data on digital 
drawings and coordinates performances on the same coordinate 
system. The target plane was divided into 5 sectors to creating 
orthographic images and those sectors were reshaped into one 
plane using reference points after image analysis. Each image 
has a different brightness according to photography conditions. 
So each image was enhanced using linear stretching for 
histogram matching. Then, different average values and 
standard deviations of digital images were regularly 
compensated on the basis of brightness data on Picture 4. Next, 
images had same color distributions by equalization. Figure 10 
displays divided sectors for creating orthographic images for 
the front plane of an object and Figure 11 presents histogram of 
digital images used. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Figure 10 Sectors Divided for Creating Orthographic Images