基于相干系数的DInSAR地表形变信息提取方法

DInSAR Surface Deformation Extraction Method Based on Coherence Coefficient

针对轨道精炼时缺少形变先验认识及地面控制点信息,盲目选择控制点可能会引入误差的问题,提出基于相干性系数的DInSAR地表形变信息提取方法。该方法通过设置相干性系数阈值剔除相干性低的像元点,以赣州市定南县作为研究区域,借助2015—2020年间的Sentinel-1A卫星数据,利用相干性系数形变信息提取方法,监测出研究区范围内-25.5~18.07 mm的地表形变信息。实验结果表明,该方法相较于常规的DInSAR方法,用于提取控制点的标准差提高了72.9%,定量表明了基于相干性系数的DInSAR方法用于选取控制点的准确性更高。

In view of the lack of prior knowledge of deformation and information of ground control points during orbit refining,blind selection of control points may introduce errors,a method for extracting surface deformation information of DInSAR based on coherence coefficient was proposed.By setting the coherence coefficient threshold value to eliminate pixel points with low coherence,Dingnan county of Ganzhou city was set as studying area,with the aid of a Sentinel-1 satellite data of between 2015 and 2020 by using coherence coefficient deformation information extraction method,and the surface deformation information was detected within the scope of studying area 25.5~18.07 mm.Simulation results quantitatively showed that it was higher accuracy for applying to select control points based on the coherence coefficient DInSAR in comparison with conventional DInSAR methods that used to extract the control points of standard deviation increased by 72.9%.