98 
GROUND DEFORMATION DETECTION ALONG THE QINGHAI-TIBET RAILWAY 
W. Y. Gong *, J. X. Zhang’ Y. H. Zhang 
Chinese Academy of Surveying and Mapping, 100039 Haidian, Beijing - (gongwy,zhangjx, yhzhang)@casm.ac.cn 
KEYWORDS: Image Processing, Earth Deformation, Interferometric SAR (InSAR), Change Detection, Ground Deformation 
Hazard Mitigation 
ABSTRACT: 
In this contribution, we focus on the ground deformation along the Qinghai-Tibet Railway based on the InSAR technique. Qinghai- 
Tibet Railway, the highest railway in the world, stretches 1,956 kilometres from Xining to Lhasa cities in western China. It lies on 
the Tibetan Plateau where the geological structural is extremely complex. And it goes through twelve regional seismotectonic zones 
where some great earthquakes had occurred. Consequently, we need to find out the ground deformation along the Qinghai-Tibet 
Railway. We have chosen a typical area located at the west of the Kunlun Mountains nearby a part of the railway where the main 
land cover types are meadow, lake and mountain. The Persistent Scatterer technique has been introduced in our experiment. It can be 
used to generate precise ground deformation in our study area. It aims at separating the phase terms (possible target motion, 
atmospheric phase contribution, decorrelation noise, residual topographic phase) by exploiting their different spectral behavior in the 
framework of a multi-dimensional joint analysis. The best place to use the PS approach is within urban area. Nevertheless, our 
researches have proven the probability to use PS technique in the Tibetan Plateau earthquake monitoring. 
1. INTRODUCTION 
1.1 Study Area 
Qinghai-Tibet Railway, the highest railway in the world, 
stretches 1,956 kilometers from Xinning to Lhasa cities in 
western China.lt lies on the Tibetan Plateau where the 
geological structural is extremely complex. It goes througth 
twelve regional seismotectonic zones where had occurred some 
great earthquakes (e.g., there was an earthquake of magnitude 
8.1 in the west to Kunlun Mountain on November 11th 2001). 
Previous research on these zones had proved that the active 
faults which have strike slop rate of 4-15 mm per year are 
possible to trigger strong earthquakes of Ms — 6-7. This factor 
and other hazard effects lead to ground deformation, which 
endanger the safety of the Qinghai-Tibet Tailway. The coupling 
of the railway and active faulting has a higher hazard to occur 
the ground deformation. 
On the other hand, such huge project in Tibet Plateau will 
certainly lead to some geologic influence that is still waiting for 
our investigation. Such research is quite important for the 
disaster forecast and environment protection. Therefore, we 
need to find out the ground deformation along the Qinghai- 
Tibet Railway. 
Interferometric synthetic aperture radar(InSAR) has proven 
capable of mapping ground deformation with centimeter-scale 
precision and producing accurate DEMs with several meters 
accuracy(Lu, 2003), and the ground coverage of thousands of 
square kilometers per image scene. 
The center latitude of the study area is 35.79 degrees and the 
longitude is 93.422. It is located at the west of part of the 
Qinghai-Tibet rail way not very far from the Kunlun Mountains 
(Fig-1)- 
The images of the study is seven segment of the ENVISAT 
image, which is 600 pixel ><600 pixel large. The resolution of 
the image is 15m in the range and 40m in the azimuth. 
Figure 1. Study area (Our study area is part of the acquired 
SAR image near the rail way) 
1.2 General Method of Research 
In this paper, we use InSAR-derived DEMs to estimate the 
deformation in the study areas. PS is a proper approach to the 
time series deformation detection, which could deal with the 
weakness of the traditional Differential InSAR processing. The 
permanent scatterers of the study area may much less than in 
the city region. The most of them may be nature’s structures, 
and they are found either in coherent and non-coherent areas. 
However, the stack of image isn’t big enough so that we have to 
do phase unwrapping before the phase model refinement. 
* Corresponding author Wenyu Gong. Chinese Academy of Surveying and Mapping, 100039 Beitaiping Road, Haidian, Beijing - 
gongwy@casm.ac.cn 
233 
■I