联合DInSAR和PIM技术的沉陷特征模拟和时序分析  

Simulation and time series analysis of mining subsidence by integrating DInSAR and PIM technology

作　　者：成晓倩[1,2] 马超[1,2] 康建荣 邹友峰[2] CHENG Xiaoqian;MA Chao;KANG Jianrong;ZOU Youfeng(School of Surveying and Land Information Engineering,H enan Polytechnic University,Jiaozuo,Henan 454000,China;Key Laboratory of Mine Spatial Information Technologies of SBSM,Henan Polytechnic University,Jiaozuo,Henan 454000,China;School of Geography,Geomatics ＆ Planning,Jiangsu Normal University,Xuzhou,Jiangsu 221116,China)

机构地区：[1]河南理工大学测绘与国土信息工程学院,河南焦作454000 [2]河南理工大学矿山空间信息技术国家测绘与地理信息局重点实验室,河南焦作454000 [3]江苏师范大学地理测绘与城乡规划学院,江苏徐州221116

出　　处：《中国矿业大学学报》2018年第5期1141-1148,共8页Journal of China University of Mining & Technology

基　　金：国家自然科学基金委员会与神华集团有限责任公司联合基金项目(U1261206)

摘　　要：高强度煤炭开采(大采高、薄基岩、快速采煤)可形成巨大的地表形变场,过大的形变相位梯度导致干涉测量失败,单独采用合成孔径雷达差分干涉测量(DInSAR)及其衍生技术都无法获得开采沉陷主值.为此,提出了联合多时相DInSAR时序分析及概率积分法(PIM),整合理论计算与卫星观测结果,实现开采沉陷特征的动态模拟和模型重构.以2012年1月—2013年6月共18期高分辨率雷达数据(RADARSAT-2,5m精细波束模式(MF5))为数据源,利用连续重访周期的DInSAR技术获得17期时间序列开采沉陷相位变化图,监测得到神东矿区布尔台矿22201-1/2工作面地表形变从产生、发展到衰退的演化规律;联合DInSAR获得的沉陷盆地边缘信息与PIM技术对矿区大变形下沉信息进行预计,两种数据整合形成混合数据集;采用Gauss函数对混合数据集进行拟合,重构矿区时序开采下沉特征曲线.研究表明:PIM技术可以弥补DInSAR技术在大形变提取上的不足,利用混合数据集建立的Gauss模型,对于有限开采(非充分采动)或充分采动的主断面下沉值具有极高的拟合度,其拟合度R2均大于0.976.A large scale deformation field can be formed by the high intensity of coal mining（large mining height,thin bedrock and fast mining）.Excessive deformation phase gradient causes interference measurement failure.It is difficult to obtain the dominant part of mining subsidence by only using Differential Interferometric Synthetic Aperture Radar（DInSAR）and its derivative technology.Therefore,the paper proposes a novel solution,which combines,multi-temporal DInSAR time series analysis and Probability Integral Method（PIM）to realize the dynamic simulation of the mining subsidence characteristics and model reconstruction by integrating the theoretical calculation and observation satellite.Eighteen set of high resolution SAR images（RADARSAT-2,5 m Multi-Fine resolution beam（MF5））acquired from January2012 to June 2013 were adopted for analysis.Firstly,continuous repeat cycle DInSAR technology to gain the seventeen time sequence mining subsidence phase variation,so as to monitor the surface deformation evolution law through generation,development and recession for high intensity mining workface 22201-1/2 of Bu’ertai mine in Shendong.Then the high reliability of subsidence basin edge information was combined with PIM technique to predict the large deformation and acquire the hybrid data set.Finally,Gauss function was adopted to fit the hybrid data set so that the reconstruction of the subsidence characteristic curve can be realized.The results show that the mining subsidence prediction of probability integral method can make up the deficiency of DInSAR technology in the extraction of large deformation.The Gauss model established by the hybrid data set has a very high degree of fitting for the dominant deformation values of finite mining（not fully mining）or fully mining.The values of R2 are all larger than 0.976.

关 键 词：开采沉陷 DINSAR 时序分析 PIM Gauss拟合模型 

分 类 号：TP028.8[自动化与计算机技术]