多尺度MT-InSAR大气延迟估计及其应用

A Multi-Scale Approach for Estimating MT-InSAR Atmospheric Delay and Its Application

大气延迟一直是制约多时相合成孔径雷达干涉测量(MT-InSAR)精度的问题之一。为抑制区域性大气窗口和对流层垂直分层效应导致的精度损失,本文引入多尺度稳健估计模型优化MT-InSAR大气延迟分量的解算和校正。实验利用MT-InSAR技术从成都主城区的14景Sentinel-1A影像中选取出高相干点,根据多尺度下对流层垂直分层延迟与高程的线性关系精确估算和校正大气延迟,提取出地表形变信息,并用同时段连续运行参考站(CORS)数据进行验证。实验结果表明:模型评估的大气延迟在0~30.2 mm,大气校正后相位均方根误差总体上减少;两类监测结果差值的均方根误差为3.9 mm,证明了本文所提出模型与方法的有效性和可靠性。

Atmospheric delay has always been the problem restricting the precision of Multi-temporal InSAR (MT-InSAR). In order to correcting vertically stratified troposphere delay, a multi-scale approach was used to determine a robust linear model. Based on the model, atmospheric correction was applied for MT-InSAR. In this research, high-coherence points were selected from 14 Sentinel-1A images of Chengdu urban areas by using MT-InSAR technology. Atmosphere delay was corrected based on the linear relationship between vertically stratified troposphere delay and elevation in multiple scales. The land deformation information was extracted and then compared with the measurements of continuously operating reference station (CORS). The result shows that atmos-phere delay values estimated by the model range from 0 to 30.2 mm. The root mean square error (RMSE) of phase is reduced in the mass after atmosphere correction. The annual mean deformation velocity of research area is about 6 mm/y, with nearly no subsidence. RMSE of difference between two measurements is 3.9 mm. Therefore, the multi-scale approach to estimating MT-InSAR atmospheric delay is proved to be effective and reliable.