COMPARISON AND IMPROVEMENT
OF POSITION METHODS OF AIRBORNE STEREO SAR IMAGES
H.D. Fan a,b , K.Z. Deng a , G.M.Huang b , Z.Zhao.\ X.J.Yue b ’ c , X.M.Luo b , Y.F.Ling ab
a. School of Environment and Spatial informatics, CUMT, Xuzhou, Jiangsu 221008, China - cumtfhd@163.com
b. Chinese Academy of Surveying and mapping. Beijing 100039, China
c. School of Geodesy and Geomatics, Wuhan University, Wuhan 430079, China
Commission VE, WG VD/6
KEY WORDS: Airborne SAR; Stereo; Algorithm; DEM; Position Model; Image
ABSTRACT: Due to the extensive application and good performance of the synthetic aperture radar (SAR), the study in SAR
image becomes more and more important. It was broadly applied in geology, ecology, hydrology, sea ice, topography map, because
of its all weather day and night. Today there are two main kinds of position methods by SAR image: interferometric SAR (InSAR)
and StereoSAR. The former can get a high precision for digital elevation model (DEM) generation, but it should meet some rigorous
conditions. However, the latter method can be achieved easily. It should only meet a certain precondition: base height ratio. So the
main feature of this method is to simplify the complex problem. Therefore, using stereoSAR to generate DEM is also playing an
important role now. In order to get better precisions, the position methods of stereo SAR images had been improved much more. To
generate DEM from stereo SAR image, there are three kinds of models mostly to be used: 1) the model of Range and Doppler
equations; 2) the equivalent collinearity equation based on the photogrammetry theory; 3) the parallax and elevation relation model
which uses the relation between parallax and elevation to calculate elevation difference and then getting the plane coordinates. In
this paper, the three kinds of position models had been introduced and compared. From that, we can find the advantages and defects
of every method, which can give us a befitting choice when we generate DEM by stereoSAR. To predigest the position methods and
improve the precisions, an improved algorithm flow was given. At last, the validity of this algorithm was proved by airborne SAR
data and better experiences results were acquired.
1. INTRODUCTION
Synthetic aperture radar (SAR) is an active microwave imaging
system . Due to its working capability in all weather and all
time, it is much more suitable for topographic mapping and
targets detective in such areas with atrocious weather.
Nowadays, there are three kinds of techniques to generate DEM
by SAR images[l]-[3]:StereoSAR or Redarclinemetry, InSAR,
Polarimetry. Although InSAR has a great advantage for its
generating accuracy, combining with some factors such as
scientific level, economy strength, application benefit and so on,
it is difficult for some countries to use InSAR technique now[4].
In our country, due to lack of SAR data, it is a little late to study
StereoSAR technique. Many researches are limited to satellite
SAR images [4] ~ [6] , but imaging model and its application of
airborne SAR are not solved very well. So, the study of this
paper has a great significance, and it can be shown as follows: 1)
StereoSAR can be fused with InSAR data and improve the
accuracy of DEM generation of InSAR technique. Researchers
in Italy after using the results of StereoSAR to correct
atmosphere parameters, the accuracy of DEM generated by
InSAR had improved from 18.14m to 13.25m [7] . 2) High
resolution is an advantage of airborne StereoSAR, and this
provides a possibility to generate a high accuracy of DEM. 3)
StereoSAR can play an important role in terrain surveying and
mapping. In some areas of west and southwest of our country,
because of raininess, cloudy and covered with forest, it is
difficult to use optical imaging system to get images. Moreover,
it is dangerous for people to do some field tasks in those areas.
As a result, the relief maps of this area are lack. So, using
airborne SAR images to generate DEM and make relief map
may be one approach to solute this problem.
2. COMPARATION OF STETEO POSITION METHODS
There are usually three kinds of methods to generate digital
elevation model by SAR image pairs [8 '^: 1) the model of
Range and Doppler equations; 2) the equivalent collinearity
equation based on the photogrammetry theory; 3) the parallax
and elevation relation model.
The model based on Range and Doppler equations and the
equivalent line central projection models based on the
photogrammetry theory have some similar characteristics. The
two methods are based on imaging model to get the three-
dimensional coordinates. The resolving steps of them can be
described as follows. Parameters of these models should be
calculated by ground control points at first, and then the ground
three-dimensional coordinates can be obtained by images
coordinates and the model parameters which have been
resolved in the first step.
The parallax and elevation relation model uses the relation
between parallax and elevation to calculate elevation difference
and then getting the plane coordinates. This method resolves
plane coordinates and elevation of the ground points separately.
It also needs to build the radar geometric model when solute the
plane coordinates. But this mode is different from the imaging
principles equations. It uses the flight attitudes of plat,
characteristics of antenna, global surface model and parameters
of map projection to build the functions of ground three-
dimensional coordinates (X,Y,Z) and images coordinates (x,y).
Comparing the three imaging models, Range-Doppler equations
which are the SAR imaging basic theories reflect the SAR
imaging principles, so the formulations are much more rigorous,
