eful. In this 
pproach for 
ur approach 
n-staggered 
n-staggered 
k for the 
1O airborne 
Berlin 
erkundung 
ytoelectronic 
mmission I, 
|. XXXI, 
3437, pp. 
Andreas Keim 
  
STRAIGHT LINES IN LINEAR ARRAY SCANNER IMAGERY 
Ayman Habib, Devin Kelley, Andinet Asmamaw 
Department of Civil and Environmental Engineering and Geodetic Science 
The Ohio State University 
habib.1 € osu.edu, kelley.83 @osu.edu, asmamaw. 1 @osu.edu 
ISPRS Commission III 
KEY WORDS: Linear Features, Linear Array Scanners, Aerial Triangulation 
ABSTRACT 
Increased use of digital imagery has facilitated the opportunity to use features, in addition to points, in photogrammetric 
applications. Straight lines are often present in object space, and prior research has focused on incorporating straight- 
line constraints into the bundle adjustment for frame imagery. In this research, we introduce a straight-line constraint 
in the bundle adjustment for linear array scanners. A linear array scanner scene is acquired using different geometry 
than frame cameras. A scene is composed of a sequence of images, each of which may be slightly shifted against each 
other due to slight changes in the system’s trajectory. As a result, straight lines in object space do not appear as 
straight lines in image space. The proposed bundle adjustment constraint accommodates this distortion. The underlying 
principal in this constraint is that the vector from the perspective center to a scene point on a straight-line feature lies on 
the plane defined by the perspective center and the two object points defining the straight line. This constraint utilizes 
the perspective transformation model for point features in linear array scanner imagery. The proposed technique makes 
use of straight-line features in object space, and aids in the recovery of the exterior orientation parameters as well as 
adding to the geometric strength of the bundle adjustment. This constraint has been embedded in a bundle adjustment 
software application, developed at the Ohio State University, that models frame and linear array scanner imagery 
1. INTRODUCTION 
Most photogrammetric applications are based on the use of distinct points. These points are often obtained through 
measurements in an analog or digital environment. Recently, more attention has been drawn to linear features. There 
are several motivations for the utilization of linear features in photogrammetry: 
* Points are not as useful as linear features when it comes to higher level tasks such as object recognition. 
* Automation of the map making process is one of the major tasks in digital photogrammetry and cartography. It is 
easier to automatically extract linear features from the imagery rather than distinct points (Kubik, 1988). 
* Images of a man made environment are rich with linear features. 
Habib (1999) discusses the various options for straight-line representation in photogrammetric applications. The 
primary representation considerations are uniqueness and singularities. In this research, straight lines in object space 
are represented by using two points along the line. In this way, the line segment is well localized in the object space. 
This representation is attractive because such coordinates can be easily introduced or obtained from a GIS database. 
Previous research on linear features in photogrammetry has focused primarily on frame imagery. Straight-line 
constraints can be incorporated into the bundle adjustment by utilizing the fact that the perspective transformation of a 
straight line is also a straight line. Mikhail and Weerawong (1994) proposed a straight-line constraint that constrains a 
unit vector defining the object space line, the vector from the perspective center to a point on the object line, and the 
Vector from the perspective center to the image point to be coplanar. In their approach, the object line is represented as 
an infinite line segment. Habib (1999) proposed a straight-line constraint, which forces the plane defined by the image 
line to be coplanar with the plane defined by the perspective center and two object points defining the line. In this 
approach the object line is defined by two points, localizing it in space. These techniques rely on the use of frame 
Imagery. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B1. Amsterdam 2000. 173