HIGH PRECISION DEM GENERATION FROM SPOT STEREO IMAGERY BY
OBJECT SPACE LEAST SQUARES MATCHING
K. C. LO , N. J. MULDER
I.T.C.
P.O.Box 6, 7500 AA Enschede
The Netherlands
(ISPRS Commission III)
Abstract:
For automatic DEM generation from multiple view SPOT imagery with high precision, an approach
of "Refinement from Coarse" is proposed. In the coarse DEM generation stage, in contrast with
the traditional methods which match at intensity level in image space with the signal
processing technique, the concept of Knowledge Engineering is used to perform high level
Feature Matching by Property List and String Matching for Correspondence analysis with high
reliability. In the refinement stage, the Object Space Least Squares Matching method, which
is the most rigorous method from a theoretical viewpoint, is proposed for coarse DEM
refinement. This method is different from the traditional two-step approach which matches the
corresponding point in image space first, and then determines the DEM by Space Intersection;
this method is improved by back mapping the image data into object space to get object
reflectance D(x,y) with referring the object surface Z(x,y), and perform matching in object
space. It is simultaneously to determine two functions in the object space: the terrain relief
Z(x,y) and the terrain reflectance D(x,y) in one solution with least squares adjustment
iteratively. The disadvantages of matching in image space which the multi view image of the
same object has different geometric or reflectance distortion can be avoided. It's flexibility
allows the user to handle more than two SPOT multi-view images in one solution, which increases
accuracy and reliability as well. It is good for SPOT images which offer the resolution with
10 meters groundel size only.
Key Words: DEM, Correspondence Analysis, Property List, String Matching,
Object Space Least Squares Matching.
I. INTRODUCTION automatic generation of coarse DEM data by Linear
Feature Matching. In order to plan and execute
The generation of a GIS for a 3-D object oriented complicated sequence of operations and functions,
data base is required urgently in many countries. we believe the methods of Knowledge Engineering
The automatic extraction of 2.5-D information from should be used, and Correspondence Analysis will
SPOT stereo images is, potentially, an efficient be based on object detection with geometric defi-
and economic way. Some on-line (real time) com- nition and object description by means of Property
mercial systems have appeared in prototype, but Lists [Mulder et al.,1988]. The selection / repre-
the methods of the off-line system for recons- sentation and use of proper knowledge is a central
tructing the earth surface with high quality and problem in research. A range of different knowl-
acceptable economy must be researched and further edge representation techniques must be developed,
developed still; We want to establish a system for along with a number of approaches to applying
generating DEM with high accuracy / high quality knowledge, which are concerned in the field of
/ high reliability with the support of the method Meta-Level Knowledge.
of Object Space Least Squares Matching. It will
then offer good fundamental 2.5-D information to 2.3 Object Space Minimum Cost Matching
GIS for multi-purpose applications.
Traditional matching in Image Space has the short-
2. BACKGROUND PROBLEMS coming that the same object in multi-view images
appears with different geometric distortions and
2.1 SPOT Satellite Imagery radiometry. The geometric distortions are caused
; by the central perspective which produces relief
Since the SPOT imagery is acquired by push-broom displacement in the image; the tilt of the sensor
scanners, the imaging geometry is different from (SPOT has off nadir angle from O to 27 degrees)
the conventional central perspective photographs causes tilt displacement in the image. As the
with frame camera, and the orbit parameters multi-view SPOT images are taken at different
require simulate dynamic modelling. For solving positions, at different times or under different
the inverse camera model problems, the point in illumination conditions, this produces different
issue is how to determine the orientation parame- radiometry for the same object in different
ters of each scan line with sufficient accuracy images. These geometric and radiometric differ-
(e.g. improving the accuracy of Tie Point measure- ences produce matching failures or reduce the
ment/transfer and best pattern of Ground Control accuracy. Therefore, motivation for improved
Point distribution) and at lowest cost (e.g. developments should come from the realization that
minimum number of Ground Control Points). On the all information in images is inherent in the
other hand, the CCT of SPOT not only presents object space, and the transformation of the
imagery data, but also offers special on board matching problem from Image Space to Object Space
auxiliary data, such as information about posi- leads to a unified and precise approach where
tion, attitude, look direction, radiometric all available knowledge is referenced to the
calibration of scene/sensor [SPOT User's Handbook, same basis. Research has to be carried out into
1988]; how to fully use this information for ways of how to model the surface of the terrain
obtaining benefits in Aerial Triangulation ( A.T.) and its reflection which is suitable for matching
and DEM generation stage must be considered in properly and efficiently.
every phase of image data processing. On the other hand, high accuracy can be obtained
by using Minimum Cost Matching; e.g. the Minimum
2.2 Knowledge Engineering Euclidean Distance can be selected as the "Cost"
in Euclidean space for similarity assessment.
The existing methods of signal correlation and Because, in case the distance is small, the dis-
feature matching are limited for handling some tance behaves as SIN function of the angle which
special correspondence analysis problems, such as is between the Feature Vectors for matching, it
133