hul 2004 
>r Urban 
tation of 
DARPA 
Region 
omputer, 
23. 
imulated 
S image 
Remote 
002/vhr/ 
. Hidden 
thophoto 
Sensing. 
t Using 
IEEE 
000, pp. 
s, M. E., 
; Using 
‘echnical 
ction of 
Mobile 
tate.edu/ 
df 
OBJECT RECOGNITION FOR INTERIOR ORIENTATION 
IN DIGITAL PHOTOGRAMMETRY 
H.Karabork *,F. Yildiz?^, E. Coskun* 
? SU, Engineering Faculty, 42031 Kampus-Konya, Turkey - (karabork, fyildiz, emrec)@selcuk.edu.tr 
ThS 12 
KEY WORDS: Digital, Photogrammetry, Object, Recognition, Aerial, Software, Matching, Orientation. 
ABSTRACT: 
In this study, measuring processes of fiducial marks are performed with area based matching method. A computer program in 
Borland C++ Builder software is developed to find out the image fiducial marks particularly for this project. Ten images were taken 
by Wild RC 10 camera. They belonged to Selcuk University Campus area. The films were scanned by Zeiss Scai Photogrammetric 
Scanner (Pixel size is 21p). Four fiducial marks on each image are automatically measured by area based matching method. Both 
residuals of fiducial marks on each image and root mean square error of transformation between pixel and image coordinate system 
are calculated. In addition, interior orientations of 10 images are manually achieved by Erdas Imagine Orthobase software. Residuals 
of fiducial marks and root mean square error of transformation between pixel and image coordinate system are manually calculated. 
Finally, both manual and automatic measurement techniques are compared with each other in terms of interior orientation accuracy. 
1. INTRODUCTION 
Image orientation is needed for any task such as the generation 
of a digital terrain model, the computation of orthophotos, and 
the acquisition of data for geographic information systems. 
Image orientation relates to the determination of parameters 
defining particular photogrammetric models for mapping 
geometric primitives such as points, lines, and areas from one 
coordinate system to another one. Image orientation has always 
been a focus on attention in photogrammetry, in view of its 
importance (Heipke® 1996). Conventionally, the measuring of 
all kinds of points, such as fiducial marks, ground control 
points, tie points and terrain elevation points, causes the most 
time-consuming and labor-intensive task in photogrammetric 
production (Lue 1999). Therefore, it is clear that automation of 
image orientation for photogrammetry is necessary. 
In this study, measuring processes of fiducial marks are 
performed with area based matching method. A computer 
program in Borland C++ Builder software is developed to find 
out the image fiducial marks, particularly for this project. 
2. INTERIOR ORIENTATION 
Interior orientation is an important task for digital 
photogrammetry. The purpose of interior orientation is to 
establish a relationship between pixel coordinate system and the 
image coordinate system (Fig.l). For digital cameras, the 
relationship between pixel and image coordinates is almost 
constant. This relationship is determined during calibration 
procedure. If digitised aerial photographs are used, this 
relationship must be determined by measuring the fiducial 
marks. Fiducial marks that have regular shapes are artificial 
objects on images. They are projected on the film at the time of 
exposure. Different cameras have different fiducial marks 
811 
(Fig.2). The fiducial marks are placed on the upper surface of 
the inside cone of an aerial camera. Many cameras have their 
own exclusive fiducial shapes (Lue 1997). They are placed 
cither at the four corners of the frame and/or in the centre of the 
four sides. Fiducial marks are registered onto the film by 
projecting an image of the mark through a small lens. Each 
fiducial is designed and projected separately (Seedahmed and 
Schenk 2000). 
  
  
  
Figure 1. Relationship between pixel coordinates system and 
image coordinates system