GENERATION OF TRUE ORTHOIMAGES FROM AIRBORNE SCANNER DATA
WITHOUT THE NEED OF HEIGHT INFORMATION
J.Albertz, D. Bomemann, H. Zoberbier
Institute for Geodesy and Geoinformation Science, Technical University Berlin, Str. des 17. Juni 135
D-10623 Berlin, Germany - albertz@igg.tu-berlin.de
Commission IV, WG IV/3
KEY WORDS: Orthoimage, True Orthophoto, Rectification, Matching, Three-Line, Aerial
ABSTRACT:
This paper presents a new technique for the efficient production of true orthoimages. The approach is based on airborne scanner data,
and takes advantage of the particular geometric properties of such data. If scanner data are available, that are acquired in two flight
lines perpendicular to each other, no height information is required to generate true orthoimages.In airborne pushbroom scanner
systems one sensor line is looking to the nadir and is adjusted perpendicular to the flight track. The data acquired by the nadir
channel follow a mixed geometry. In flight direction the mapped area is imaged in parallel projection. Across to the flight line the
terrain is imaged in central projection. In other words, in flight direction, object points recorded are not displaced due to their height,
but mapped in a correct position. On the other hand point displacements because of object heights occur only crosswise to the flight
directions scene is imaged twice in flight lines perpendicular to each other, the first image strip provides correct ground coordinates
of any object in one direction, the second image strip in the other direction. The new approach takes advantage of these particular
geometric properties. It is based on the definition of corresponding points in the image data sets by means of matching techniques.
Each conjugate point directly provides the correct ground coordinates and can be assigned as a pixel to the orthoimage. Thus, a true
orthoimage is generated without any knowledge of the object heights or further calculations. The paper provides a detailed
description of the approach, and its advantages and limitations are discussed.
1. INTRODUCTION
Traditionally differential rectification of aerial photographs is
applied to generate orthoimages. This requires a height model
to correct for the radial displacements, which are caused by the
central perspectivity of the images. It is obvious that the quality
of the product depends on the precision of the height model
provided. It is common to use Digital Terrain Models (DTM)
describing the topographical surface of the terrain. However,
these models do not contain objects above the terrain surface,
such as buildings, bridges etc. Therefore such objects cannot be
corrected and rectified to their true position in the reference
plane. In order to achieve improved results, objects above the
terrain must be modeled in a Digital Surface Model (DSM).
Utilizing such models it is possible to derive orthoimages with
high objects in their correct position, thus the result is a "True
orthoimage".
Many approaches have been developed to generate DSMs, e.g.
from urban areas. But in any case it is a very time-consuming
and cost-intensive procedure to generate such height models. In
other words to provide these data remains the bottle-neck of
true orthoimage generation.
In this paper we introduce a new approach for generating true
orthoimages. It is based on digital airborne data, recorded with
line-scanning cameras - so-called Pushbroom Scanners - and
does not require any information about the height or geometry
of the objects. Also, the paper presents the first results of recent
research with real data from a line-scanning camera, the pre
processing of data, the assumptions and the achievable
geometric accuracies.
The following section of this paper describes briefly the
traditional approach of producing true orthoimages, its con
ditions, limitations and drawbacks.
2. THE TRADITIONAL APPROACH FOR THE
PRODUCTION OF TRUE ORTHOIMAGES
It is common to generate orthoimages from aerial photo-graphs
or from data acquired by digital frame cameras. Due to the
central projection the acquired aerial photographs contain scale
variations and especially radial relief displacements. In the
process of differential rectification, digital elevation models are
used to consider the effects of relief displacements. Therefore,
orthoimages represent the terrain in parallel projection with a
uniform scale and the relief in its true ground position.
However, objects above the relief surface or reference plane,
like buildings, bridges, trees, etc. are usually not described in
the used elevation model. Therefore such objects, mainly man
made objects, are displaced from their true ground position. The
orthoimage is partly geometrically inaccurate and shows
leaning buildings and bent bridges (e.g. Mayr, 2002). Some
important ground features like footpath, streets and other
objects are hidden for the user of the orthoimage. By applying
the differential rectification approach for large-scale
orthoimages over urban areas, also double mapped areas may
appear. In other words, these orthoimages do not have the same
characteristics as a map and it can be difficult to locate objects,
compute areas, measure distances and obtain other useful
information (Habib et al., 2007). The mentioned shortcomings
make it difficult to integrate orthoimages as important
components into a GIS database.