The international Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
181 
For the same amount of ground deformation, the induced 
DInSAR phase fringes will be denser in the results generated 
using shorter radar wavelength (e.g. C-band) than the ones 
generated using longer wavelength (e.g. L-band). Figure 2 (a) 
shows the high phase gradient within the subsidence bowl 
causing the phase ambiguous. As a result, the phase fringes 
cannot be fully recovered, even after being filtered. 
Figure 1. Histograms of coherence values of (a) - (c) ALOS 
interferometric pairs: 14/6/2007 - 30/7/2007 Bperp = 515m, 
30/7/2007 - 14/9/2007 Bperp = 143m, and 15/12/2007 - 
30/01/2008 Bperp = 351m, and (d) ENVISAT pair: 3/1/2008 - 
7/2/2008 Bperp = 361m, respectively. The mean coherence 
value of these 4 pairs are 0.355, 0.461, 0.586 and 0.379. 
5. RESULTS AND DISCUSSIONS 
The study area is about 24km x 25km. The differential 
interferograms of the selected four interferometric pairs (3 from 
ALOS and 1 from ENVISAT) are shown in Figure 3. In the 
results, over 21 mine subsidence spots are identified using 
DInSAR technique. The magnitude of ground deformation due 
to underground mining can be more accurately measured or 
recovered if the interferometric pair illustrates high coherence 
over the scene and at the same time the deformation phases are 
not ambiguous. In this study, the ALOS pair with the mean 
coherence value of 0.586 shows the deformation fringes quite 
clearly as shown in Figure 2 (b). 
Figure 4 shows the final deformation map derived from the 
ALOS pair, 15/12/2007 - 30/01/2008. The result is overlaid on 
a high resolution optical image with the aid of geographic 
information systems (GIS). The maximum subsidence detected 
due to active underground mining activities over a period of 46 
days is about 36cm. 
There are several coal collieries in the local area and they are 
closely located to each other. Even though the proposed mining 
location and mining schedule are recorded, it is still hard to 
track the precise locations of mining activities. But it is 
extremely important for ensuring the safety of the miners as 
well as the local residents when the mining progress at the two 
collieries are close to each other. The results of DInSAR 
derived from successive acquisitions were post-analysed against 
other supporting spatial data in geographic information systems 
(GIS). By comparing the locations against the proposed mining 
schedule, any illegal mining activities can also be identified. 
6. CONCLUDING REMARKS 
This paper demonstrated the capability of differential radar 
interferometry (DInSAR) for monitoring the ground surface 
deformation caused by underground mining activities. The 
study area chosen in this paper is in the Northern China. 
ENVISAT and ALOS SAR data acquired in C- and L-band 
respectively were tested in this paper. Within an area of 
approximately 24km x 25km, more than 20 subsidence spots 
are identified using DInSAR. The results show ALOS data are 
more suitable for mine subsidence monitoring. The L-band 
radar signal and the high imaging resolution (8m) of ALOS 
reduce the problems caused by high deformation phase gradient 
and improve the scene coherence. It is also found that the local 
agricultural areas may introduce noises into the interferometric 
results, especially during summer season. Nevertheless, the area 
of impact of mine subsidence at multiple collieries can be 
monitored at the same time using DInSAR. The DInSAR 
deformation map is complementary to the ground surveys and 
assists to identify any abnormal subsidence behaviour. Together 
with other spatial data, it helps to ensure the safety of the local 
communities as well as the mine workers. 
Figure 2. Enlarged DInSAR phase fringes induced by ground 
deformation of (a) ENVISAT (C-band) pair: 3/1/2008 - 
7/2/2008 Bperp = 361m, and (b) ALOS (L-band) pair: 
15/12/2007 - 30/01/2008 Bperp = 351m.