基于小波域HMT模型InSAR干涉图噪声滤波研究

相干斑噪声是InSAR干涉图固有的,且InSAR干涉图对相干斑噪声十分敏感,为了得到精度更高的干涉测量产品,需要对干涉图进行相干斑噪声滤波。针对InSAR干涉图中相干斑噪声的统计特性,提出了一种基于小波域HMT模型的InSAR干涉图滤波算法,对InSAR干涉图的实部和虚部分别进行处理。试验研究结果表明,该方法在有效抑制相干斑的同时,还能有效地保持相位的细节信息和条纹的边缘结构,而且大大地减少了残余点的数量,有利于提高InSAR干涉测量的精度。     

新型多基线DInSAR地表形变监测技术研究动态

合成孔径雷达干涉测量（InSAR）是一项广泛采用的雷达遥感测量技术,可以获取大区域、长时间、毫米级的地表形变监测,是SAR图像应用研究的热点。从InSAR技术监测地表形变时面临的问题出发,分析了近年来多基线DInSAR方法的新进展,论述了相干目标算法\,分布目标算法及SAR层析成像技术等在监测地表运动时的原理及技术应用,详细讨论了DInSAR地表形变监测由二维参数研究发展至三维、四维空间,由城区发展至广阔非城区地表监测的发展趋势。     

融合PS、SBAS、DS InSAR技术的昆明地面沉降研究

时序InSAR技术为城市地面沉降监测和防治提供了有效方法,然而PS-InSAR和SBASInSAR技术自身的缺陷限制了InSAR技术的监测精度,特别是复杂地形产生的低相干性引起的PS点稀疏问题。在PS和SBAS算法基础上,通过引入Stacking技术联合PS-InSAR进行相干目标点选取,提出融合PS、SBAS、DS-InSAR的技术,并对比提出的方法与常规PS+SBAS-InSAR的监测结果。研究结果表明：提出的方法与PS+SBAS-InSAR反演的昆明沉降速率结果存在较好的一致性,其中本实验提出的方法能增强观测区域点目标空间分布的密度,进而得到更多有效的地表形变信息。从整个研究区来看,昆明市城区地表整体存在-22～8 mm/a的沉降速率,严重沉降区集中在官渡区、西山区和五华区,并已经形成多个沉降漏斗。自1989年以来,小板桥和河尾村依然是最严重的两个沉降漏斗中心,而东北方向的蒋家营则是本次研究中发现的新沉降点。结合历史资料的验证分析表明：昆明地面沉降主要受地下水抽汲、建筑荷载与工程施工以及断陷盆地的构造运动的影响。     

多模式雷达在矿区沉降监测中的应用研究

合成孔径雷达差分干涉测量(DInSAR)技术在地表形变监测方面已得到广泛应用。介绍了将差分InSAR技术运用于矿区地表沉降监测,获得了河北峰峰煤矿地表Envisat/ASAR和ALOS/PALSAR的雷达形变干涉相位图,并对Envisat C波段和ALOS L波段的形变干涉相位图进行了相干特性和相位特性的分析。通过综合考虑C波段和L波段的优势与不足,将两者联合使用,实验表明利用多模式雷达数据对矿区地表沉降进行检测的可行性。同时,通过对雷达干涉相位图的分析,能够及时提供正在进行地下开采活动的矿区地理位置。     

极具应用潜力的PS技术

虽然在一些条件良好的地区,合成孔径雷达差分干涉测量(D-InSAR)技术已经取得了不少令人瞩目的应用成果,但时间失相干因素很大地限制了D-InSAR的广泛应用,而大气效应也会影响D-InSAR的测量精度。近年来提出的永久散射(Permanent Scatterers, PS)技术则在很大程度上解决了这两个难题,即使在无法获得干涉条纹的情况下,利用基于多时相SAR图像和相位稳定像元点集的PS技术也能取得毫米级的地表形变运动测量精度,因而PS技术大大增强了干涉测量的环境适应能力及其精度,是干涉研究领域的一项重大技术突破,具有巨大的实际应用潜力。首先着重介绍PS技术的原理及其处理过程,然后对PS技术的应用,以及PS技术与传统D-InSAR技术之间的关系进行扼要的讨论,期望有助于推动这项高新技术手段在国内的研究和应用。     

基于多时相干涉SAR提取地表形变的关键技术探究

基于重复轨道的合成孔径雷达差分干涉测量(D-InSAR,Differential Interferometric SAR)技术自问世以来,就被立即成为研究重点并在许多领域得到应用。但是常规D-InSAR技术易受时间、空间失相干与大气延迟等影响而有时无法得到可靠的形变信息,因此难以在实践中得到常态化应用。永久散射体技术与小基线集技术的出现,有效地克服了上述问题的制约。本文融合了两种技术的优点,介绍了多时相干涉SAR技术的关键处理步骤,并基于多时相干涉SAR技术提取了太原市2005～2009年间的形变信息。     

雷达InSAR数据在矿山地表形变监测中的应用研究

针对矿山地表形变这一特殊区域，开展星载InSAR测量的研究工作，在大量调研的基础上确定技术流程和关键技术及解决方法，选择具有典型意义的河北武安地区，应用InSAP数据，采用三次、二次过境差分法处理，获得了几个时间间隔内矿山开采地表形变图像。经与实际对比应用效果较好。真实地反映了地表形变的动态信息。实践证明，利用InSAR数据进行地表形变动态监测有着重要的现实意义和广泛的应用前景。     

重轨星载InSAR测量中的大气校正方法综述

随着合成孔径雷达干涉测量(Interferometric Synthetic Aperture Radar,InSAR)技术在地形测绘及地表形变监测等领域应用日益广泛,提高其测量精度成为该技术领域近期研究的热点之一。而InSAR测量过程中的大气效应是限制其测量精度提高的主要误差源之一,如何经济有效地消除大气效应引起的延迟位相(即大气校正)是InSAR技术应用中亟待解决的问题。详细介绍了InSAR测量中现有的大气校正方法,并评述了各自的优缺点及适用条件。最后对InSAR大气校正研究中仍就存在的问题进行了概括并提出展望。     

星载InSAR非相干图像对精确配准

为了解决星载InSAR非相干图像配准中不同区域对应不同形变参数的问题,将几种传统的SAR图像配准和光学图像配准的方法有机结合,提出了一种基于概略配准、Harris角点检测、相关系数匹配、Delaunay三角网剖分插值映射的InSAR复图像精确配准方法。通过仿真验证,该方法能够适用相干和非相干的InSAR复图像。配准精度达到0.1像素级。     

顾及相干系数的InSAR形变场采样算法

利用SAR差分干涉测量技术获取的大范围连续覆盖地表形变场为断层活动性质、发震断层参数等反演研究提供了丰富的数据信息。为提高形变机制反演数据源的可靠性和结果的准确性,提出一种顾及InSAR相干系数的概率采样算法。该算法基于四叉树数据压缩模型,将窗口形变梯度以及自适应核运算后的相干表征量按权重分配作为分割窗口的约束参数,以此建立概率采样函数;并引入数字高程模型对无效采样信息进行掩膜处理,达到兼顾提高采样点质量和保留细节形变信息的目的。通过对2016年中国台湾地区美浓M_w6.7级地震同震形变场数据进行实验分析,结果表明,顾及相干系数的概率采样算法在保留高质量数据和重要细节信息方面优势更为明显。     

机载X-波段双天线InSAR数据森林树高估测方法

短波长的干涉合成孔径雷达(InSAR)适用于数字表面模型(DSM)提取,但难以提取准确的林下地相位,在缺乏高精度数字高程模型(DEM)的森林区域,短波长InSAR数据估测树高的能力受到限制。针对这一问题,采用机载X-波段单极化(HH)双天线InSAR数据开展了森林树高估测方法研究。双天线InSAR可以忽略时间去相干的影响,并且X-波段波长较短,入射角较大(中心入射角45.77°),地表对干涉去相干的贡献可以忽略,因此可将干涉复相干作为体去相干,对体去相干模型中的结构函数进行勒让德展开,截取第0阶展开式得到了基于相干幅度的森林树高估测模型,利用均匀选取的LiDAR冠层高度模型(CHM)检验样本对估测结果进行严格的精度评价,并与差分法的树高估测结果进行对比。精度评价结果显示:相干幅度法与差分法都得到了较高的估测精度,两者的R~2、RMSE、总精度分别为0.81、0.86;1.20m、0.97m;86.4%、88.7%。研究结果表明:相干幅度与森林树高具有负相关关系,适用于估测树高,基于单极化相干幅度的估测模型也可以得到较高的估测精度,与差分法的估测结果相比,虽然估测精度略有降低,但此方法具有两方面的优势:一方面,估测结果不需要实测样地数据标定,对于没有实测样地数据的森林区域亦能进行高精度的树高估测;另一方面,相干幅度法不需要高精度的DEM,具有更强的实用性。     

COSMO-SkyMed数据在常州市地表形变监测中的应用

小基线集合成孔径雷达干涉测量技术(SBAS-InSAR)已成功应用于城市地表形变监测,并表现出极大的潜力和优势。X波段高分辨率雷达卫星在地表微小形变探测方面较C波段和L波段更为敏感。选取覆盖常州地区COSMO-SkyMed高分辨率SAR影像,采用SBAS-InSAR方法获得了地表形变时间序列,对比水准观测数据,分析了干涉测量结果的精度,根据历史地下水位监测数据,分析了地下水水位变化对地表形变的影响。结果表明:干涉测量结果与水准观测数据具有很好的一致性,沉降区域主要发生在武进区,最大沉降量超过-40mm,主城区出现了轻微的回弹现象,回弹达到+5mm;地下水水位持续上升与地面沉降减缓、地面回弹趋势一致,地下水水位变化仍然是常州市地表形变的主要影响因素。     

A survey of temporal decorrelation from spaceborne L-Band repeat-pass InSAR

In this paper we quantify the effects of temporal decorrelation in repeat pass synthetic aperture radar interferometry (InSAR). Temporal decorrelation causes significant uncertainties in vegetation parameter estimates obtained using various InSAR techniques, which are desired on a global scale. Because of its stochastic nature temporal decorrelation is hard to model and isolate. In this paper we analyze temporal decorrelation statistically as observed in a large swath of SIR-C L-Band InSAR data collected over the eastern United States, with a repeat pass duration of one day in October 1994 and a near zero perpendicular baseline. The very small baseline for this particular pair makes the effect of volumetric scattering on correlation magnitude statistics nearly imperceptible, allowing for a quantitative analysis of temporal effects alone. The swath analyzed in this paper spans more than a million hectares of terrain comprised primarily of deciduous and evergreen forests, agricultural land, water and urban areas. The relationships of these different land-cover types, phenology and weather conditions (i.e. precipitation and wind) on the measures of interferometric correlation is analyzed in what amounts to be the most geographically extensive analysis of this phenomenon to date.     

Optimal estimation of interferometric phase for measuring surface deformation

In recent times, time-series interferometric synthetic aperture radar (InSAR) methods are developed to retrieve the deformation signal in non-urban areas from distributed scatterers (DS). Phase triangulation algorithm (PTA), an important step in these methods for filtering decorrelation noise from DS, aims at optimal estimation of the filtered wrapped interferometric phase values using InSAR data stack. The uniqueness of this research work lies in the incorporation of one such PTA only to provide an optimal set of wrapped interferometric phase values before phase unwrapping in the open source StaMPS processing environment. The proposed methodology, when adapted to measure surface deformation in Tehri reservoir rim region, Uttarakhand, India using Environmental Satellite (Envisat) C-band advanced synthetic aperture radar images, works efficiently and has enhanced the spatial coverage of measurement pixels compared to standalone PS-InSAR processing. It is also revealed from the one-dimension-line of sight velocity map that resulted velocity estimates are congruent with standalone PS-InSAR processing.     

The present status of subsiding land vulnerable to roof collapse in the Jharia Coalfield,India, as obtained from shorter temporal baseline C-band DInSAR by smaller spatial subset unwrapped phase profiling

In this paper, we identified recently subsiding areas in Jharia Coalfield, Jharkhand, India from the shorter temporal baseline Radarsat-2 C-band interferometric synthetic aperture radar (InSAR) data pairs of 2012. Although shorter wavelength C-band differential InSAR (DInSAR) is more sensitive to slow deformation and better suited for higher precision land subsidence measurement, the dynamic and adverse land cover in mining areas and resulting temporal decorrelation problem poses a serious problem for DInSAR observation in mining areas. We used smaller temporal baseline data pairs and adopted InSAR coherence-guided incremental filtering with smaller moving windows to highlight the deformation fringes over temporal decorrelation noise. We identified the deformation fringes and validated them based on ground information to prepare the land subsidence map of the coalfield in 2012. Several new, previously unreported subsidence areas were detected in the present study with a total subsiding area of 6.9 km². The recent incidence of roof collapse on 15 November 2014 at Angar Patra village in Katras region of the coalfield where 45 houses collapsed and 10 people were injured is situated in a highly subsiding vulnerable area as obtained from the present study. Due to spatial discontinuities of InSAR coherence, DInSAR phase unwrapping for the entire study area in one go did not appear feasible. To avoid this problem, we performed DInSAR processing in smaller spatial subsets and unwrapping of the subset interferograms by a ‘minimum cost flow’ algorithm. Subsequently, we plotted unwrapped phase profiles across the deformation fringes and retrieved the maximum deformation phase with respect to background phase and translated them into radar line of sight (LOS) displacement rates. For obtaining the average subsidence rates, we adopted InSAR coherence-weighted LOS displacement rates taking into account the contribution of each data pair as a function of DInSAR phase quality of the fringe areas. Ground-based subsidence measurements by precision levelling were conducted in four test sites that had been undergoing active underground mining during the observation period. We compared space-borne DInSAR-based subsidence rates obtained by the adopted technique with precision levelling measurements. Overall, the results are found to agree well. In the four test sites with gentle to flat topography, land subsidence occurs at slow to moderate rates due to compression of in-filled material (resulting from sand stowing in underground mining), without any evidence of roof collapse. In such cases, the horizontal displacement component is less significant, and overall surface displacement occurs essentially in the vertical direction. However, we assessed the nature of subtle horizontal strain to infer relative shrinkage or dilation of the land surface which could be additive or subtractive to vertical displacement in DInSAR-based LOS displacement.     

星载SAR干涉技术获取DEM及其精度分析

星载合成孔径雷达干涉（InSAR）技术是一种数据覆盖范围广、廉价、高效、方便的数字高程模型（DEM）获取方法,但在地面植被覆盖广、大气水汽含量高的地区其影像相干性随时间基线的增加迅速降低;同时,SAR卫星的轨道误差也影响DEM精度。利用ERS-1/2卫星串行模式SAR数据获取镇江地区DEM,分析了轨道误差对DEM精度的影响;根据干涉相位的统计特性,从理论上给出干涉相位噪声与相干系数和视数之间的关系。实验结果表明就干涉像对的卫星轨道误差和相位噪声而言,在小区域内DEM精度优于3.5 m。     

Monitoring of large-scale deformation in mining areas using sub-band InSAR and the probability integral fusion method

The large-scale and rapid land subsidence that occurs in mining areas often leads to problems, such as densely spaced interference fringes and the temporal decorrelation of interferometric synthetic aperture radar (InSAR) interferograms. To solve these problems, sub-band InSAR is applied to monitor the large-scale deformation that occurs in mining areas. First of all, four different bandwidth images with three sub-band bandwidth parameters are used to extract simulated mining-induced subsidence with seven different deformation magnitudes. The results of the simulation experiment suggest the following conclusions. In monitoring subsidence with different deformation magnitudes using images with different bandwidths, an optimal monitoring value exists; wider image bandwidths lead to smaller optimal monitoring values and higher monitoring accuracies. Therefore, an appropriate sub-band bandwidth should be selected that depends upon the image bandwidth and the subsidence level to achieve optimal monitoring. The optimal sub-band bandwidth for monitoring subsidence of different magnitudes in mining areas is determined through simulation experiments, and these conclusions can provide a technical basis for selecting the appropriate sub-band bandwidth for the monitoring of subsidence in mining areas. Although sub-band InSAR can reduce the number of interference fringes and the difficulty of unwrapping, the simultaneous introduction of large amounts of noise leads to reduced monitoring precision, and the application of the probability integral method in the prediction of mine subsidence is more mature. Therefore, the combined use of sub-band InSAR and the probability integral fusion method to monitor mining-induced deformation is proposed in this paper. The probability integral method is used to perform noise peeling on the interferometric phases of the sub-bands to improve the monitoring accuracy of sub-band interferometry. Then, according to the results of the simulation experiment, the fusion method with the appropriate sub-band bandwidth parameters is applied to monitor the surface deformation associated with working face 52,304 from 2 December 2012 to 13 December 2012. Finally, the monitoring results are compared with the results of monitoring using conventional differential interferometric synthetic aperture radar (D-InSAR) and global positioning system (GPS) field survey data. The results show that the reliability and accuracy of the fusion method are much better than those of conventional D-InSAR in monitoring the large-scale deformation that occurs at the edges of subsidence basins.     

基线误差、相位误差和大气延迟误差对InSAR数据处理的影响分析

基于星载InSAR的基本原理,首先推导了基线误差、相位误差和大气延迟误差对高程测量的影响公式,得出了他们均与有效基线长度成反比的影响规律。接着讨论了大气延迟误差对二轨法和三轨法形变测量的影响规律:大气延迟误差对二轨法形变测量的影响随入射角变化不明显,对三轨法形变测量的影响随基线比变化不明显。当入射角取23°,基线比取1/2,欲使大气延迟误差对形变测量的影响＜1 cm,二轨法要求其测量误差＜6.5 mm,三轨法要求其测量误差＜7.5 mm。     

Simulation of time-series surface deformation to validate a multi-interferogram InSAR processing technique

A refined, small baseline subset (SBAS) interferometric synthetic aperture (InSAR) method has been developed to estimate time-series surface deformation through fusion of multi-interferogram processing. Using a synthetic data set that takes into account two time-varying deformation sources, topography-induced errors, atmospheric delay anomalies, orbital errors and temporal decorrelation, all of which are based on realistic ERS-1/ERS-2 SAR image acquisition dates and baseline configuration over the Seguam volcano, Alaska, we assess the accuracy of the refined SBAS technique. Detailed comparison between SBAS-derived products, including time-series deformation maps, atmospheric delays and baseline errors with those of synthetic values, attest the robustness of the refined SBAS technique. The root mean square error of the mean deformation rate between the simulated and SBAS-retrieved is about 0.66 mm year^–1. Thus, the measurement accuracy of the refined SBAS method would be around 1 mm year^–1.