基于岭估计形变模型解算的矿区地面沉降SBAS−InSAR监测及分析

    SBAS-InSAR monitoring and analysis of land subsidence in mining areas based on the solution of ridge-estimated deformation models

    • 摘要: 为提高基于SBAS−InSAR技术的矿区地面沉降监测精度,基于Sentinel−1A影像数据,以济宁市某矿区为研究区,采用岭估计方法对所构建的SBAS−InSAR形变模型进行解算,并通过水准测量数据对岭估计(RR)和最小二乘估计(LS)2种方法解算得到的结果进行精度验证和对比分析。结果表明:2种方法得到的矿区地面沉降时空演变规律一致,但监测到的最大累积沉降量分别为−128.1、−227.8 mm,存在一定差异;RR方法的数据拟合效果显著优于LS方法,二者的平均内符合精度分别为2.77、77.66 mm;RR方法解算结果与水准数据之间的相对误差较小,平均RMSE为10.5 mm,时间序列监测结果与水准数据吻合一致,LS方法解算结果精度较低,平均RMSE为21.7 mm,时间序列沉降趋势波动较大。

       

      Abstract: In order to enhance the accuracy of ground subsidence monitoring in mining areas using SBAS-InSAR technology, based on the Sentinel-1A image data, a certain mining area in Jining City was selected as the research area. The ridge estimation method was used to calculate the SBAS-InSAR deformation model constructed. The accuracy of the results obtained by the ridge regression (RR) estimation and the least squares (LS) estimation methods was verified and compared through the leveling measurement data. The results show that the temporal and spatial evolution patterns of ground subsidence in the mining area obtained by the two methods are consistent, but the maximum cumulative subsidence amounts measured are −128.1 mm and −227.8 mm respectively, indicating a certain difference; the data fitting effect of the RR method is significantly better than that of the LS method, and the average internal consistency accuracies of the two methods are 2.77 mm and 77.66 mm respectively; the relative error between the solution results of the RR method and the leveling data is small, with an average RMSE of 10.5 mm; the time series monitoring results are consistent with the leveling data, while the solution accuracy of the LS method is lower, with an average RMSE of 21.7 mm, and the fluctuation of the time series subsidence trend is relatively large.

       

    /

    返回文章
    返回