基于多元回归模型的GB-SAR监测误差改正及形变分析

针对基于地基合成孔径雷达(GB-SAR)露天矿边坡监测原始影像干涉相位误差改正不准确导致形变数据精度偏低的关键问题，以马兰庄露天铁矿GB-SAR原始影像为数据源，对原始影像干涉相位误差来源和分布特征进行了分析.提出了利用三重阈值提取PS点和高质量PS点的方法，并基于高质量PS点相位与像元坐标建立了多元回归模型.依据该模型对PS点进行相位改正以获取准确形变相位和单幅影像形变量，并在时间序列上进行叠加分析.研究结果表明，基于多元回归模型的GB-SAR形变监测误差改正方法可以准确改正GB-SAR形变监测误差，提高了形变监测数据的精度.

Error Correction and Deformation Analysis of GB-SAR Monitoring Based on the Multiple Regression Model

Aiming at the key problem of low precision of the deformation data due to the inaccurate correction of the original image interference phase error based on the ground-based synthetic aperture radar(GB-SAR) in open pit slope， the sources and distribution characteristics of interference phase error in original image were analyzed with the original GB-SAR image of Malanzhuang open-pit iron mine as the data source. The method of extracting PS points and high-quality PS points by using triple threshold was also proposed. In addition， a multi-regression model was established based on the high-quality PS point phase and pixel coordinates. According to the model， the PS point phase was corrected to obtain the accurate deformation phase and the single image deformation， and the superposition deformation analysis was performed on the time series. The results show that the GB-SAR deformation monitoring error correction method based on multiple regression model can accurately correct the GB-SAR deformation monitoring error and improve the accuracy of deformation monitoring data.