| Coal is one the major sources of energy of China. The problems of ground subsidence and environmental hazards caused by high intensity and large areas exploiting coal resources become more and more serious. In order to reduce mining subsidence disasters, we should carry out a continuous and efficient monitoring on mining subsidence and obtain the rules of ground subsidence and the degree of damage, which help us make plans of coal mining and control ground subsidence disasters. Traditional monitoring methods require large amounts of human and material resources and carry the disadvantage of being based on point-wise measurements and lacking area information, which can not meet the production needs of development. The appearance of interferometric synthetic aperture radar (InSAR) technology provides a new method for ground subsidence monitoring. Due to its outstanding advantages of fast, high-precision, all-time, all-weather and large areas, it makes up for the deficiencies of the traditional monitoring methods to a large extent. However, InSAR technology has many limits when applying in mining subsidence monitoring due to the special characteristics of geographical environment of coal mining areas and the complexity of mining subsidence (rapid, large gradient deformation). Moreover, atmospheric effect, temporal and spatial decorrelation also influence the imaging quality. Therefore, this paper focuses on the problems that exist in mining subsidence monitoring when using InSAR technology and its relevant algorithms. The main contributions are described as follows.(1) The research status of InSAR technique were summarized. The insufficiency of InSAR and its related techniques studies were pointed out. The basic principles of SAR, InSAR, D-InSAR and time series InSAR technology were described, and several important parameters of InSAR system were also analyzed, including inteferometric phase, inteferometric coherence and the sensitivity of interferometric phase to elevation change.(2) A new automatic registration approach based on a multi-step matching strategy was proposed. In the first step, key points were detected and matched using an improved scale invariant feature transform (SIFT) operator. In the second step, region correlation matching (RCM) algorithm was used to exclude matched points with low correlations. In the third step, random sample consensus (RANSAC) algorithm was used to conduct purifving of matched points once again and finallv an exact match points were obtained. The proposed method does not any priori information, such as satellite orbit information and external DEM. The experiment results show that this method could meet the requirement of high accuracy registration at different conditions, including different wavelengths, different spatial resolutions, different correlations.(3) A method was proposed that integrates InSAR data and point cloud data obtained by terrestrial LiDAR to improve the detectable deformation gradient of InSAR technology. The proposed method takes advantage of high-density of terrestrial LiDAR point cloud data and its high precision of point positioning after 3D modeling. Based on inverse distance weighting (IDW) algorithm, on the one hand, large gradient deformation and decorrelation area existing in InSAR deformation field was filled with point data, on the other hand, the deformation field of common coverage area monitored by InSAR and terrestrial LiDAR was taken a weighted average to generate a higher accuracy monitoring results. The proposed method can solve the problem of large gradient deformation to some extent and it provides new ways and means for obtaining large gradient deformation monitoring by InSAR technology in mining area.(4) A time series deformation model was proposed that combines SAR image phase measurement and amplitude measurement to monitor time series mining subsidence. On the one hand, this method used the amplitude-based feature tracking (ABFT) technology to monitor large gradient deformation, on the other hand it used interferometric phase measurement to conduct monitoring in small deformation area, followed by the integrating of the two deformation results. Compared with merely using phase measurement technology, the integrated results can obtain the deformation information in the large gradient deformation area. Likewise, compared with merely using intensity tracking technology, the integrated results can get more accurate small deformation information.(5) A new ultrashort baseline (USB) InSAR method was introduced to monitor the subsidence of old goaf. Compared to the traditional differential interferometric technology, it has the unique advantage, i.e. with no need for external DEM,which avoids the introduction of new errors. Using the method obtained subsidence velocity and time series subsidence results of old goaf, moreover, the relationship between cyclic period of subsidence velocity and H/M/c and relationship between fluctuant peak of subsidence velocity and mining thick were established, which provides the basis for the prediction and evaluation of old goaf residual deformation.(6) A model was proposed that integrates InSAR technology and the support vector regression (SVR) algorithm to monitor and dynamically predict mining subsidence. InSAR technology was first used to monitor the range of influence and the development trend of mining subsidence, thus obtaining the law of surface subsidence. Based on the monitoring results obtained by InSAR technology, the SVR algorithm was used to describe the nonlinear function correlativity between the monitored data and future subsidence. As the performance of the SVR algorithm depends largely on the choice of relevant parameters, the particle swarm optimization (PSO) algorithm was introduced to select the optimal parameters for the SVR algorithm. Finally, a method of rolling prediction based on the optimized SVR parameters was adopted to update the training and learning samples of SVR, thus allowing the algorithm to use the latest monitored data to dynamically predict future mining subsidence. |
PDF Full Text Request