Study On D-InSAR Monitoring Mining Subsidence Base On Offset-Tracking Method

Using traditional means such as leveling and GPS technology to monitor coal mining subsidence can get real time and high precision results, but there are also some disadvantages such as big workload, limited monitoring scope, having no access to a large area of monitoring results. By contrast, rapid developing D-In SAR technology using SAR images because of the day, all-weather, wide coverage, strong penetrating power and other characteristics, can quickly obtain a large area of subsidence basin and the deformation field. The precision can reach millimeter, it has obtained the good application in coal mining subsidence monitoring. However, complex geological conditions, large magnitude and fast speed of mining subsidence, limit the applications of D-In SAR technique in coal mining subsidence monitoring. The domestic and foreign experts and scholars have done a lot of research and exploration. The offset-tracking technology develops and extends the application scope of D-In SAR technology, which mainly used in the earthquake, glaciers, landslides and other aspects, and has made a series of achievements. They gradually begin to study its application in coal mining subsidence monitoring, but there are many aspects to be further research.D-In SAR technology uses the phase information of SAR image. In this paper, it uses offset-tracking technology through the amplitude information of SAR image, and then calculates the mining subsidence. D-In SAR technology is mainly affected by image loss of coherence and large deformation. Offset-tracking technology can solve the problem of the gradient magnitude of large deformation monitoring, and do not need phase unwrapping at the same time, so it’s not affected by coherence.To this point, the research content and the results in this paper are as follows:(1)Through the experiments, it analyses the influence of vegetation factors on image coherence and explores the influence of monitoring level on surface subsidence information interpretation. The results show that vegetation coverage is the decisive factors influencing the coherence between images; vegetation sparse in winter is more suitable than in the summer for using D-In SAR technique for the surface deformation monitoring. Thus it is difficult to obtain slow mining surface deformation for a long time; In view of the mining ground subsidence laws, it determines the scope of application between D-In SAR technique maximum monitoring magnitude and the actual mining conditions.(2)Through algorithms and GAMMA software realizes offset tracking method to get the subpixel offset, which adopts the registration method based on the search window to achieve subpixel registration. Implementation process involves a series of command parameters; the optimization of parameters is obtained by repeatedly experiments. Bilinear function model and offset statistical method in the absence of surface deformation area is constructed in this paper. Mining subsidence is obtained combining with the actual image data. Compared with the result of D-In SAR qualitatively and quantitatively, the fitted values by bilinear function model have better removal effect and its subsidence changes is closer to the actual situation.(3) When converting the offset to 3D deformation, the assumption method is used to calculate the vertical variables directly under ignoring the influence of the horizontal movement. In order to find the vector relationship between horizontal deformation and vertical deformation, the model of probability integral method is established considering the influence of horizontal movement. Affected by the horizontal movement coefficient and the direction, the change of the deformation is very large. With reference to the results by D-In SAR monitoring, the results show that: the probability integral method and the assumption method can monitor the subsidence center and are in agreement with the D-In SAR results; The assumption method’s subsidence basin shape is closer to the situation of D-In SAR basin; The two methods can successfully extract a large number of monitoring magnitude value, reaching to meter level; When interpolating D-In SAR missing values, there is a serious fault and a large gap.(4) Choosing Radarsat-2 images as the data source and Jiulong in Fengfeng mining area as the study area, the ground deformation in the mining area is calculated by migration algorithm and GAMMA. And compared with the leveling measurement results, the research finds that the sinking curve of migration algorithm is closer to the observatory sinking curve trend, but has great difference. The GAMMA subsidence curve is very different from the observatory sinking curve. However, the partial observation station and the measured value is less than 100 mm.