针对永久散射体差分干涉测量(PSInSAR)算法流程,发展了基于周期函数模型的空间维解缠方法,并将其应用于矿区时间序列地表变形反演。通过在研究区域内安装人工角反射器(CR),将CR点上计算所得的周期模型参数分量作为整个网络的约束,通过空间约束平差以实现空间维解缠。选取了河南省境内白沙水库附近的煤矿密集区为主要研究区域,采用周期函数模型对矿区线性及非线性形变分量进行模拟,反演了2007年2月-2010年2月的时间序列形变场,并采用研究区域内的水准实测数据作为外部验证数据。实验结果表明:白沙水库周围区域存在着较为明显的沉降,在煤矿分布区域内累积最大沉降量超过了10 cm。沉降区域内以线性沉降趋势为主,非线性沉降较为缓慢,仅在水库的西南方向较为明显。应用已有的水准点实测形变值对实验结果进行验证分析,结果表明该方法精度可达约±2.1 mm,证实了本文采用的方法在矿区地表时序形变反演中的可行性和可靠性,对预防过度采矿导致的矿区塌陷具重要现实意义。

Due to the frequent mining activity and increasing mine geological disasters in our country, the long term dynamical monitoring and analysis of the mining area are of great importance to prevent the potential geological damage in mining area. The Permanent Scatterer Interferometric Synthetic Aperture Radar (PSInSAR), a newly developed ground deformation monitoring technique, which may not be influenced greatly by the spatial and temporal deccorelation, has been widely applied in study on regional displacement, including the deformation of urban area, terrain area and mining area. During the step of spatial unwrapping in PSInSAR algorithm, stable points or external known GPS points are necessary. The procedure of selecting the reference stable points is obviously uncertain and the external GPS data is difficult to obtain.
Due to the shortcomings of traditional spatial unwrapping in PSInSAR algorithm, a new method of spatial unwrapping based on periodic function is developed. Since Corner Reflector (CR) point can be installed easily, which can be applied in the area without external constrained data (such as GPS data) and avoid the uncertainty of choosing reference point in the PS parametric adjustment network, the subsidence rates calculated on CR points are used as constraints for PS network while the spatial unwrapping is performed using the parametric adjustment method. With the improved method, the PSInSAR is applied in the inversion of time series ground deformation in mining area. With 14 ALOS PALSAR images from February 2007 to February 2010, the deformation inversion experiment is carried out. The colliery dense distribution area, around Baisha reservoir in Henan province, is chosen as the study area in the experiment. 6546 PS points except CR points are detected during the experiments. The linear velocities calculated out through traditional spatial unwrapping method are compared to that of the developed method.
The algorithm achieves the integration of CR data and PSInSAR algorithm for the first time. The authors succeed to inverse the time series of subsidence from February in 2007 to February in 2010, using the periodic function to simulate the linear and nonlinear components of the deformation. The results show that there appears obviously time series subsidence around the reservoir, with the max value over 10 cm in the colliery distribution area, due to mining activities. The subsidence mainly performs to be linear subsidence. The nonlinear subsidence only appears to be a little obvious in the northeast of the reservoir. In order to validate the result of the experiment, deformation monitoring with leveling was also carried out in the area. With comparison to the deformation result of leveling, the accuracy of ± 2.1mm is calculated. It can be concluded from the good accordance that the method has the following advantages:(1) CR point can be installed easily, hence we can choose the study area freely, which can be applied in the area without external constrained data (such as GPS data); (2) CR Point can be taken as constraining data for the spatial unwrapping of PS network and increase the redundancy number of parametric adjustment model which can make the solutions more stable; (3) Corner reflectors have high reflectivity which can be identified easily on the SAR image, thus the inaccuracy within the step of coordinate transformation can be avoided and the accuracy of the solutions can be improved. (4) It can avoid the uncertainty of choosing reference point in the PS parametric adjustment network.