ANALYSIS OF LAND SUBSIDENCE MONITORING IN MINING AREA WITH TIME-SERIES INSAR TECHNOLOGY

Time-series InSAR technology has become a popular land subsidence monitoring method in recent years, because of its advantages such as high accuracy, wide area, low expenditure, intensive monitoring points and free from accessibility restrictions. In this paper, we applied two kinds of satellite data, ALOS PALSAR and RADARSAT-2, to get the subsidence monitoring results of the study area in two time periods by time-series InSAR technology. By analyzing the deformation range, rate and amount, the time-series analysis of land subsidence in mining area was realized. The results show that InSAR technology could be used to monitor land subsidence in large area and meet the demand of subsidence monitoring in mining area.


INTRODUCTION
It is well known that the mining of underground mineral resources can easily produce ground subsidence.The larger mining amount and the longer mining time make the subsidence more obvious.Serious land subsidence will form collapse pits, ground fissures, and some even form large area of water accumulation pits, which threaten the safety of life and property of local residents and bring huge economic losses.
The traditional monitoring method of mining subsidence is Leveling and GPS, etc. Due to the sparse measurement points and long interval of retest time, these methods can hardly meet the requirements of large area continuous measurements.And these methods have defects, such as wasting time, energy and financial.
In addition, the measurement points are difficult to protect (Wang et al., 2014;Liu et al., 2005).Since the idea of synthetic aperture radar interferometry was first proposed in 1974, InSAR technology has developed rapidly.It provides a new technical means for land subsidence monitoring, because of it has the ability of working all-weather situation and all-day, etc.
References (Hayman et al., 2010;Jiang et al., 2011;Yin et al., 2011) have realized the subsidence monitoring of mining area by time-series InSAR technology.Although there are still some problems to be improved, but there is no denying that with the in-depth study of InSAR technology, the realization of engineering monitoring is the trend of the future.This paper takes a mining area as the experimental area, and uses time-series InSAR technology to monitor the land subsidence.This paper realized the time-series analysis of the deformation results by using two satellite images, ALOS PALSAR and RadarSat-2, time spans nearly 10 years.

Time-series InSAR Technology
References (Ferretti et al., 1999;Mora et al., 2001;Zhang et al.,2009;Zhang et al.,2012) have analyzed and applied the timeseries InSAR technology by different ways, so not tired in words here.In this paper, GDEMSI and Gamma software are used to realize the data processing, including radar data preprocessing, stable points extraction, Delaunay triangulation establishment, linear deformation estimation, separation and estimation of nonlinear deformation with atmospheric phase, total deformation estimation, etc.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3, 2018 ISPRS TC III Mid-term Symposium "Developments, Technologies and Applications in Remote Sensing", 7-10 May, Beijing, China

Data Selection
According to previous experience, data selection mainly considers these factors, as follows: (1) The images completely cover the survey area.The dates of data are continuous and the number of them are sufficient.Then make sure the data fit the time-series analysis.
(2) The wavelength of the data is long enough, and the penetration is strong, which is suitable for the large deformation monitoring in the mining area.
Through data analysis, two kinds of data are selected in this paper: The first one is ALOS-PALSAR data in L-band, from January 2007 to January 2011, 23 scenes; The other one is RadarSat-2 data in C-band, from March 2015 to May 2017, 32 scenes.Table 1 and Table 2

Data Processing
After data preprocessing, extract the stable highly coherent point

Figure8. The location of deformation area
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3, 2018 ISPRS TC III Mid-term Symposium "Developments, Technologies and Applications in Remote Sensing", 7-10 May, Beijing, China

Accuracy Verification
The first stage data is archived data, and there is no comparable leveling data in this period from 2007 to 2011, so its accuracy could not be verified.The second stage data is the programming data, during the data collection period, second-class leveling measurement is carried out.By comparing the accuracy of the leveling data with the monitoring results, the mean error of the rate meets the requirements of the technical design.

Explanation Before Analysis
Since two stages data are involved, so two periods results will be explained next, including the influence range, deformation rate and deformation amount.But in fact, the results of the two periods could not be directly compared, because their reference datum are different.The first one is based on the situation in 2007, and the other is based on the situation in 2015.Therefore, the comparison in values is meaningless, but it can reflect the trends.
So, the values discussed below are talking about the changes in the trends of the both.

Range Analysis
According to the method in 2.4, we get the edge of each deformation area.Here we count the acreage of each area, and analysis the changes between the two period data.Table 3 is the statistical result of the deformation range, it can be seen from the table that the changes in the deformation range of the two periods are prodigious.The range of deformation area 4# has been obviously reduced, after verification, it was found that the mine had stopped mining.The total area of the two periods increased from 1281.75hm 2 in 4 years to 1326.41hm 2 in 2 years.There is a big influence in a shorter time.

Rate Analysis
In this paper, the deformation rate of two periods was analyzed, and table 4 is the statistical results.In the first stage, the maximum deformation rate reached -192 mm/a in deformation area 2#.Besides, the maximum rate of deformation area 3# and deformation area 6# exceeds -100mm/a.In the second stage, the maximum deformation rate is in deformation area 7#, the maximum deformation rate of deformation area 2# and deformation area 3# are -114.49mm/a and -85.02 mm/a.
Although the results of two periods could not be compared directly, but we can see that they have changed a lot in value.

CONCLUSIONS
Because the monitoring project is limited by some reasons, more detailed description of data processing and analysis results could not be made.But through analysis, it can be seen that: ( (2)Although the two stage data could not be compared directly, but we can still do some easy deformation trend analysis via value.
(3)Except mining area, InSAR technology can also be used to monitor subway, highway, railway and building accurately.
Therefore, this method has a broad application space in the future.
target, then calculate relative linear change rate and relative elevation error of each two points via model coherence coefficient.A highly coherent point with known variables and DEM error is chosen as the reference point.Integrate the relative linear change rate and relative elevation error of each two highly coherent points, and then get the linear deformation rate and height error of each point target.Finally, separate the nonlinear deformation phase and atmospheric influence phase base on residual phase, then obtain the time-series deformation results according to the linear and nonlinear deformation phase.The first stage data is processed by GDEMSI software, time baseline is 650 d, spatial baseline is 1400 m, and 62 interference pairs are generated.Figure 1 is the small baseline spatiotemporal distribution of the first stage data.The second stage data is processed by Gamma software, time baseline is 120 d, space baseline is 212 m, and 69 interference pairs are generated.Figure 2 is the small baseline spatiotemporal distribution of the second stage data.

Figure 1 .
Figure 1.The small baseline spatiotemporal distribution of the first stage data

Figure 7 .
Figure 7. Dividing method of the deformation range According to the deformation range above, we found 7 subsidence areas, which distributes in a line from southwest to 1)InSAR technology can monitor ground subsidence in large areas.It can reflect the subsidence phenomenon better, and meet the demand of subsidence monitoring in mining area.Using InSAR time-series technology to process the data can obtain the time-series change of the monitoring area, which is convenient for the subsequent subsidence analysis.If we can realize dynamic monitoring and share the results with other departments, it could provide a reference data for mine monitoring and management.
are the baseline information of the data.

Table 3 .
Statistical results of deformation range

Table 4
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3, 2018 ISPRS TC III Mid-term Symposium "Developments, Technologies and Applications in Remote Sensing", 7-10 May, Beijing, China

Table 5 .
Statistical results of the maximum value of cumulated variables unit：mm

Table 6 .
Statistical results of themaximum value of the annual variable unit：mm