Mingsheng Liao 
  
AUTOMATIC REGISTRATION FOR REPEAT-TRACK INSAR DATA PROCESSING 
Mingsheng LIAO, Li ZHANG, Zuxun ZHANG, Jiangqing ZHANG 
Wuhan Technical University of Surveying and Mapping, 
Natinal Lab. for Information Eng. in Surveying, Mapping and Remote Sensing 
Luoyu Road 129, Wuhan 430079, P. R. China 
liao? wtusm.edu.cn 
Working Group IC-1 
KEY WORDS: InSAR, Registration, SAR processing, Automatic algorithm Remote Sensing. 
ABSTRACT 
Interferometric Synthetic Aperture Radar (InSAR) allows production of high resolution DEM and detection of small 
earth motions using multiple pass SAR data sets obtained by remote sensing satellite. But the digital processing to 
extract the DEM is quite complicated. The whole procedure has not yet reached sufficient efficiency and robustness to 
warrant automated DEM production as commonly produced by stereo vision with optical images. The automatic 
algorithm for precision registration is one of the bottlenecks for improving both the accuracy and the efficiency of 
multi-source data analysis, including the repeat-track D-InSAR processing. In this paper, an automatic approach with 
multi-step image matching algorithm is presented. All procedures could be automatically implemented. The primary 
experiment result is promising in the fast precision registration for the repeat-track InSAR data and reveals the potential 
of the presented automatic strategy. 
1 INTRODUCTION 
InSAR presents a completely new approach for topographic mapping. But the whole process has not yet reached 
sufficient robustness to warrant automated DEM production as commonly produced by stereo vision with optical image 
(Trouve 1998). Moreover, the computing time should be kept as short as possible for an operative data processing 
system. One of the critical factors in generating the interferograms is the efficient and precise registration of the single 
look complex data. To obtain a high quality interference fringe pattern, it is necessary for the images to be registered to 
sub-pixel accuracy (Gabriel 1988, Zebker 1986). 
Usually image registration deals with geometrical transformation and resampling of the pixel values. The identical 
points on both images can be identified as the control points. If sufficient control point pairs are determined, the 
transformation function could be chosen. The new pixel values are calculated according to the geometric transformation 
and the resample formulae bi-linear or bi-cubic interpolation). The determination of suitable control point pairs has 
become the critical point. 
The traditional method for registration usually utilizes the tie points which are manually determined. However, the 
manual procedure constitutes a tedious and time-consuming task and may be impossible for SLC SAR data. Various 
researchers have investigated the automatic matching approach (Zhang 1998). But it is much more difficult in SAR 
image matching than in traditional optical image since the speckle and the blur texture feature. 
Lin et al presented the registration algorithm based on the optimization of a target with the average fluctuation function 
of the phase difference image (Lin 1992). Gabriel ea t| presented an approach based on the coherence estimation, which 
was adopted by many investigators (Gabriel 1988). A two-dimensional fast Fourier transform was performed on the 
interference fringe generated in the matching window, then the relative quality was assessed with the signal-noise ratio 
of the fringe spectrum. Zebker et al also used small patches as tie points on ERS images and performed the 2D 
convolutions to find maximum power spectra. Offsets in range and azimuth can be determined based on the difference 
of these patches (Zebker 1994). But in fact, the power distribution will differ greatly from the ideal case with a single 
peak. Multiple peaks may appear, resulting in difficulty in correlation determination (Wang 1990). In these algorithms, 
the generation of the fringe and the FFT are the very time consuming task. Additionally the entire matching procedure 
is repeated as the search area was scanned pixel by pixel. 
In this paper, a multi-step matching approach based on intensity or power detected images is presented. In the primary 
matching, the distinct points are selected and their conjugation points are determined automatically based on the 
+ This research was funded by National Nature Science Foundation of China (Contract No. 69782001) and R&D Foundation for Surveying 
and Mapping Technology (Contract No99008 ). 
  
186 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part Bl. Amsterdam 2000. 
Th 
tw 
fe: 
ba 
He 
the 
ad 
the 
It i 
ov 
sec 
22 
As 
COI 
goi 
anc 
alg 
Sin 
ma