Research On Multi-source Data Fusion Technology And Its Application In Mining Subsidence Monitoring

Due to the gradual reduction of coal resources in the east,the key mining area has gradually shifted to the west where the ecological environment is more fragile in recent years.The underground coal resources in the western region are characterized by shallow burial depth,thick coal seam,low hardness of surrounding rock,and thick sandy topsoil.During large-scale and stressful mining,it is easy to induce severe subsidence on the surface,resulting in large cracks and soil texture.Problems such as loosening,vegetation destruction,and changes in soil moisture may eventually cause surface ecological damage.In the previous research,observation stations are established to carried out the observation of mining subsidence in most major mining areas in China.Reliable observation data and a large number of observation line monitoring data are obtained,which provided abundant field data to support the study of subsidence caused by underground mining.However,it is difficult for observation stations to be completely located in the feature point areas in the western mine with complex topography because of the complex terrain and gaps in the western sand dune mining area.What’s more,observation in the scene could be different and inefficient,the traditional method of obtaining the subsidence data from the surface deformation observation station gradually exposes some shortcomings.Proceeds from formulating a scheme for combining the Satellite-space-ground observation technologies,optimizing the solution method,and combining high-resolution multi-spectral remote sensing images,this thesis mainly studies the construction of high-efficiency and high-precision subsidence basins,the evaluation and application of surface ecology and inversion of mining subsidence prediction parameters.The main conclusions are as follows.(1)With the gradual transfer of coal mining resources to the west,the topography,topography and geological conditions of the mining areas in the central and western regions are quite different.In order to observe the surface subsidence in a complete time series,the observation method should be improved.The method of combining the sky and the ground is used for observation,and different technical combination schemes are formulated according to the actual situation and the characteristics of the Satellite-space-ground observation technologies.(2)By using the law of error propagation and standard deviation model,the high-precision DEM obtained by UAV photogrammetry and the ALOS DEM are integrated,so that the accuracy of the external DEM required for SAR image calculation can be increased to the decimeter level,and then the accuracy of SAR image resolution and the ability to identify deformation can be improved.(3)The field data acquisition and processing scheme of UAV mobile photogrammetry has been improved,and a high-brightness image control point device has been designed.Experiments showed that the best image control point layout is based on evenly layout all around,point group layout in four corner,and increasing density layout in subsidence basin.The observation data of GNSS image control points is modified by DIn SAR deformation data to improve the coordinate accuracy of image control points,and then the accuracy of the aerial triangulation is improved which made the DEM RMSE increased by 0.011 m and subsidence RMSE increased by 0.012 m.(4)For the processing of spatial 3D point cloud data obtained by 3D laser scanning,data registration and filtering algorithms have been improved.Combining the Bursa model and the least square method,the 3D point cloud data is converted from relative coordinates to absolute coordinates,that is,each site is converted into geodetic coordinates,so that there is no target in the monitoring areas.At the same time,the adaptive progressive triangulation encryption algorithm is improved,the grid is divided according to the terrain relief,and the threshold of the number of point clouds in the grid is set.The obtained ground points are used as seed points,the seed point data is integrated into the progressive triangulation network algorithm,the point cloud data and the normal distance and angle of the lowest point are calculated,and the ground points are finally obtained,and then the time series DEM is constructed.(5)According to the observation data obtained by the combination scheme of the Satellite-space-ground observation technologies,the coordinate system is transformed into Beijing 54 coordinate system,and the high-precision data is extracted by combining the data and experimental precision of the above three theories.On the premise of maximizing the use of high-precision data,inverse distance interpolation,local polynomial interpolation and empirical Bayes Kriging interpolation are used to fuse the high-precision data,then the accuracy of the data obtained by the three fusion methods is compared.Thus,which can be determined is that the best scheme is the combination of Satellite-space-ground observation technologies,and the fusion accuracy of empirical Bayes Kriging interpolation is the best fusion method with an accuracy of 0.038 m.Experiments show that using the combination of the Satellite-space-ground observation technologies to construct a whole subsidence basin and then inverse the subsidence parameters,the results can be the most accurate.Using the subsidence parameters to predict probability integral,the median error accounts for 4.3% of the maximum subsidence value,and the relative error is far less than 10%,which meets the accuracy requirement of surface subsidence prediction.(6)The optical remote sensing image is used to retrieve the surface ecological index of the western sand dune mining area,and the principal component analysis is used to calculate the surface ecological index,and the influence of the subsidence area on the surface ecology is analyzed from the subsidence monitoring results.In conclusion,basing on the theories and experiments in this thesis,the research programme and method on constructing subsidence basin and inverting the probability integral are reliable,which is meaningful to the evaluation and management of surface ecological damage.There are 51 diagrams,29 tables and 232 references in this thesis.