【作者】 王帆；

【导师】 徐世浙； 张登荣；

【作者基本信息】 浙江大学 ， 地球探测与信息技术， 2011， 博士

【摘要】 星载合成孔径雷达干涉测量技术(InSAR)是一种可以全天候、全天时大范围快速获取地表地形和形变信息的具有巨大发展潜力的新型卫星遥感对地观测技术。在我国,由于中西部地区地形复杂,人类难以涉足,人工地面测量较为困难；东部工农业发达,各种人类活动导致灾害性地表形变现象时常发生,而使用人工长期监测的方式不仅成本高昂,而且空间分辨率不高。相对于常规人工地面测量方式受工作区、气候和成本等条件局限,InSAR干涉测量技术具有很大的优势。然而,受多种因素的影响,在实际应用中InSAR干涉测量常常难以达到理论精度。这些影响因素包括地表散射体物理性质随时间变化而导致的时间失相干,空间基线误差导致的几何去相干,而星载SAR两次观测期间因大气状态变化而导致的相位延迟也使得测量精度产生误差。其他因素还包括轨道误差、地形误差和系统热噪声等。这些因素都限制了InSAR干涉测量的具体应用。本文结合点目标提取分析的思想,在InSAR干涉测量获取地表地形和时序形变信息领域进行了深入的研究。在提取地表地形方面,利用相干性系数提取的高相干点目标,结合外部DEM数据,建立多元线性回归模型,可以有效削减空间相关的大气相位屏效应(APS)和基线误差等造成的InSAR相位干扰,提高所获取的数字高程精度。在提取时序地表形变方面,为了探究研究区域InSAR干涉数据集在时序上受APS的影响,尝试利用外部大气数据获取各个时间节点之间的大气相位延迟场,分析了大气效应在InSAR相位中的贡献。为了从一定程度上避开误差严重的低相干区,研究从数学和物理意义上提取长时间序列SAR数据集的相位稳定点目标(相干性好、信噪比高的点目标),首先利用一种基于小波变换的自适应梯度滤波方法用于提高干涉图集的信噪比,然后利用幅度离散系数法快速获取候选点目标集(PSC),再利用各类相关因子,包括负相关系数、信噪比和概率分析等对PSC点集进行筛选,最终进行解缠和时空滤波分析提取形变信息。这一过程借鉴了斯坦福大学的StaMPS方法,统一建立了干涉模型和相位模型,利用时空统计模型进行定量分析提取地表形变信息。通过对实验结果的验证和分析,获取了研究区域的亚厘米级地表形变数据和整体态势,并在此基础上结合地质、水文地质和人为活动等影响因素,对研究区进行探索性的地表形变易发性分析。更多还原

【Abstract】 Spaceborne synthetic aperture radar interferometry (InSAR) is a new remote sensing technology which can obtain large areas of information on surface topography and deformation all-time and all-weather. The terrain in central and western China is too complex to get involved for humans, so it’s difficult to do any artificial ground-based measurements.In eastern China, the industrial and agricultural is well developed, human activities lead to frequent disasters of all kinds of surface deformation. The cost of long-term monitoring of artificial is high and its spatial resolution is low. So the conventional measurements are limited by the ground-work area, climate conditions and costs while InSAR technology has great advantages. However, in the practical application of InSAR technology, it is often difficult to achieve the theoretical accuracy for some factors:changes in the physical property of surface scatters over time lead to decoherence; baseline error leads to geometric decoherence; the effect of atmospheric phase screen leads to interferometry measurement error.The other factors include orbit errors, topography error, system thermal noise and so on. These factors have limited the specific application of InSAR technology.In this paper, we study the InSAR technology for obtaining the information of surface topography and the timing deformation with the extraction and analysis of the point targets. In the study of the extraction of surface topography, to improve the accuracy of digital elevation model(DEM) obtained by InSAR technology, we estimate the coherence map of the interferogram to extract high-coherence point targets and establish multiple linear regression model with external DEM data to reduce InSAR phase errors caused by the spatial correlated effects of atmospheric phase screen (APS) and baseline error. In the study of the extraction of the timing surface deformation, we use the external atmospheric data to obtain the regional atmospheric phase delay field between different SAR observation times to study the APS effect on InSAR phase. To avoid errors from decoherence areas, the phase stable point targets(high coherence and high SNR) are extracted in mathematical and physical sense from SAR data sets in long time series. Firstly we use an adaptivea gradient filter method based on wavelet transformation to improve signal to noise ratio of interferogram data sets. Secondly the amplitude dispersion index is used to select the candidate persistent scatterers(PSC) quickly. Then several factors such as negative correlation coefficient, SNR and probability analysis factor are used to do the selection of PSs from PSC. Finally the interferograms are unwrapped and spatial-temperal filtered to extract the surface deformation. We set this process to establish an interferometric phase mode and utilize the spatial-temperal model to do the quantitative analysis by leaning from Stanford method for persistent scatterers(StaMPS). In the study area, surface deformation of sub-centimeter accuracy and the overall situation are obtained by the result of experiments. We also analyze the surface deformation susceptibility in the study area based on several factors such as geology, hydrogeology, human activities and so on.更多还原