251 
APPLICATION OF DINSAR AND GIS FOR UNDERGROUND MINE SUBSIDENCE 
MONITORING 
MIAO Fang, YAN Mingxing, QI Xiaoying, YE Chengming, WANG Baocun, LIU Rui, CHEN Jianhua 
Chengdu University of Technology 610059, Sichuan, China 
mf@cdut.edu.cn yanmingxing@gmail.com 
KEY WORDS: DInSAR; GIS; Mine subsidence; ERS; Envisat; Tangshan 
ABSTRACT: 
This research used both ERS and Envisat images to investigate the feasibility of differential radar interferometry (DInSAR) for mine 
subsidence monitoring in Tang Shan, Hebei Province, China. DInSAR results are analyzed and validated with the aid of Geographic 
Information System (GIS) tools. The drawbacks of using interferometric measurements for mine subsidence monitoring are discussed. 
The repeat-pass tandem and Envisat DInSAR results are presented. 
1. INTRODUCTION 
Interferometric synthetic aperture radar (InSAR) systems 
exploit the phase differences between two SAR images acquired 
over the same area. Useful topographic information, such as 
digital elevation model (DEM), can be derived. Differential 
InSAR (DInSAR) has been further used to measure the 
deformation of the ground terrain. A number of experiments 
have demonstrated that InSAR is very useful in such fields as 
earthquake-related deformation, volcanic motion, ice-sheet shift, 
urban settlement. [Massonet and Feigl, 1998; Rosen et al., 
2000] 
Compared with the conventional approaches (such as GPS 
monitoring), using of InSAR and D-InSAR in surface 
deformation monitoring can cover a large area on the earth, and 
the result can be obtained in a relatively short time. The cost of 
InSAR is lower and it is very useful for the rural area or the 
dangerous places where we can’t easily arrive. Finally, because 
the cloud and the light have no effect on the Synthetic Aperture 
Radar images, the images can be obtained every times when the 
SAR satellites pass the area. The feasibility and capability of 
DInSAR for underground mine subsidence monitoring have 
been tested in the UK [Wright, P. and R. Stow. 1999], France 
[Carnee, C. and C. Delacourt, 2000], Germany [Wegmuller, 
2000]. In these studies, the images acquired by the two ERS 
satellites are the only data source. 
This paper reports the progress of the ongoing ESA CAT-1 
project (ID 4527). We used radar images acquired by the ERS 
and Envisat satellites to investigate the use of radar 
interferometry for mining-induced subsidence monitoring in 
Tang Shan, Hebei Province, China. Successful DInSAR results 
are exported to the GIS and mine subsidence regions extracted. 
The DInSAR results are analyzed and validated against other 
spatial information, such as TM images and mine plans. 
2. REPEAT-PASS DINSAR 
Repeat-pass space-borne DInSAR has been used to derive 
ground displacement maps. Two SAR images acquired from 
two slightly different positions, at different revisit times, are 
used to measure the phase difference, or so-called interferogram, 
between the two acquisitions. 
As shown in (l)[Liu Guo-xiang,2006], the phase change in the 
interferogram is the composite of systematic phase(also termed 
flat-earth trend phases)from the reference surface,' 4> fiat, 
topographic information, <i> topo , surface displacement between 
the two acquisitions, 4> disp , atmospheric delay, 4> de ia y , and noise, 
4* noise- 
4* — 4> flat - * - 4* topo 4> disp"*" 4> delay 4* noise (0 
DInSAR requires the removal of phase signatures that are 
contributed by the flat-earth and topography, and so isolating 
the ground displacement component. The 4> fi at can be predicted 
using the satellite state vectors or baseline data and based on the 
interferometric geometry, and then subtracted from the initial 
interferogram. The 4> top o can be simulated and eliminated by 
introducing DEM information. The atmospheric component, 4> 
delay > is primarily due to fluctuations of water content in the 
atmosphere between the satellite and the ground, it is difficult to 
eliminate because the absence of the weather-information and 
the limited resolution of the SAR sensors. We can use filter to 
enhance the signal-to-noise level. 
3. STUDY AREA AND USED DATA 
Tangshan City and Kailuan Mining Area, located in east of 
China, are selected as the experimental district.Tangshan city is 
the main coal city in China. Since 1970s, underground mining 
extended to downtown area. Especially since 1990s, 
underground mining has induced large area of land surface 
subsidence; many buildings, road and pubic establishment were 
damaged. Kailuan Mining Area has been exploited for 123 years, 
the mining area covers 670 km 2 , and the subsidence area, 
affected by underground coal mining, covers 208 km 2 .[Wu 
Lixing,2005] 
Figure 1 is TM image of Tangshan test site. 4 ERS images 
spanning from 1996 to 2000 and 3 ENVISAT spanning from 
2004 to 2006, were combined to produce interferograms 
(table l&table2). Two different approaches were applied to 
construct the differential interferogram: three-pass method using