HILLY 
25 of 
Vol. 29, 
Fundamental Analytic of Satellite CCD Camera Imagery Using Affine Transformation 
Tetsu Ono, 
Kyoto University 
Atsushi Okamoto, Kyoto University 
Susumu Hattori, 
Fukuyama University 
Hiroyuki Hasegawa, PASCO Corporation 
COMMISSION III 
KEY WORDS: Satellite CCD Camera Imagery, Affine Transformation 
ABSTRACT 
An orientation theory of two-dimensional affine imagery was first presented by Okamoto (1992), which can be 
applied for ultra-precise measurement of very small objects. This theory may be employed for the analysis of 
satellite CCD camera imagery by clarifying the geometrical characteristics of the model connection problem with 
adjacent stereo models and transforming the central-perspective images into affine ones. Therefore, in this paper, the 
orientation and model connection problems of two-dimensional affine images are first discussed in detail. Then, the 
transformation of central-perspective images into affine ones is explained with correction of the image 
transformation errors due to height differences in the terrain. The proposed method was applied for space 
triangulation of simulated satellite CCD camera images taken consecutively and proved to have a fairly high 
accuracy. 
INTRODUCTION 
In satellite photogrammetry, imaging devices (CCD 
camera and CCD line-scanner) have usually very narrow 
field angle. Therefore, we encounter great difficulty to 
apply the conventional orientation theory based on 
projective transformation for the analysis of satellite 
imagery. In order to overcome this problem, we should 
develop another orientation theory. Affine 
transformation may be a very promising one, because 
affine transformation pertains to parallel projection and 
thus the flying height of the satellite plays no role in 
the geometry of an affine image. 
The orientation theory of two-dimensional affine 
images was first derived by Okamoto in 1992. In order 
to employ this theory for the analysis of satellite CCD 
camera imagery, the central-perspective imagery must 
be transformed into an affine one by using the 
approximations of the orientation parameters. This 
image transformation cannot be carried out without 
errors due to both deviations of the approximations and 
height differences in the terrain. Thus, a correction 
method of the image transformation errors is required. 
In this paper, the orientation problem of 
two-dimensional affine imagery is extended to the 
model connection with adjacent stereo models so as to 
discuss space triangulation of satellite CCD camera 
imagery based on affine transformation. Then, the 
transformation of central-perspective images into affine 
ones is described in detail and the correction of the 
image transformation errors is discussed, in which they 
are classified into two types: errors due to deviations of 
the approximations of the orientation parameters and 
errors caused by height differences in the terrain. The 
611 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
methods proposed here are tested with simulated 
examples so as to explore the difficulties when 
applying them to practical cases. 
GENERAL ORIENTATION THEORY OF TWO-DI- 
MENSIONAL AFFINE IMAGES 
ORIENTATION PROBLEM OF A STEREOPAIR OF 
AFFINE IMAGES 
Let a three-dimensional object space (X, Y, Z) be 
projected into a plane based on affine transformation 
(See Figure-1.). The basic equations relating an object 
    
  
measured image plane 
(affine image) 
object space 
X 
Figure-1 : parallel projection of an object space 
into the measured plane of the compa- 
rator coordinate system 
LAB es CESSE ES i PE : 4