Corner reflectors as the tie between InSAR and GNSS measurements: Case study of resource extraction in Australia
Corner reflectors as the tie between InSAR and GNSS measurements: Case study of resource extraction in Australia
复制标题
角反射器作为 InSAR 和 GNSS 测量之间的纽带：澳大利亚资源开采案例研究
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
M. Thankappan
中科院分区：
文献类型：
--
作者：
M. Garthwaite;S. Lawrie;J. Dawson;M. Thankappan
覆盖澳大利亚的稀疏点网络上的连续全球导航卫星系统 (GNSS) 测量与干涉合成孔径雷达 (InSAR) 的频率相对较低但空间密度较高的观测相结合，是为澳大利亚开发的新大地测量参考系的基础。认识到提高定位精度的经济重要性以及大地测量工具有助于了解能源相关问题的潜力，澳大利亚政府资助了一个由 GNSS 测量标记和共置雷达角反射器组成的创新区域大地测量网络。该新网络已安装在昆士兰州苏拉特盆地，由于从地下开采大量资源，预计该地区会出现区域沉降。在本文中，我们提出了对苏拉特盆地 TerraSAR-X、Sentinel-1A、RADARSAT-2 和 ALOS-2 SAR 图像中角反射器响应得出的先验视线高度误差的初步观测结果。 1. GNSS 和 INSAR 的组合 永久部署的角反射器允许从 SAR 采集中提取精确的地面运动，作为已知位置处具有稳定相位特性的持久散射体 [1, 2]。由于包含角反射器的分辨率单元的信噪比远大于包含自然散射体的分辨率单元的信噪比，因此持久散射体信号可用于验证使用传统差分 InSAR 时间序列技术在 SAR 场景中空间测量的变形信号 [3, 4]。角反射器与 GNSS 测量标记的共置将确保 SAR 导出的变形产品可以与当地大地测量参考系联系起来（并有助于通过变形模型进行约束）。 2. 苏拉特盆地大地测量网络 澳大利亚对能源需求的不断增长导致非常规煤层气储量的开采增加，特别是在昆士兰州的苏拉特和加利利盆地。苏拉特盆地于 2006 年开始生产煤层气，目前有多家运营商正在生产储量。基于多孔弹性模型和气体生产率的沉降幅度预测表明可能正在发生分米量级的沉降[5]。由于预期信号波长约为 100 公里，我们在昆士兰州苏拉特盆地北部部署了一个新的永久大地测量网络（图 1）[6]。该网络由 65 个 GNSS 测量标记站点组成，其中 40 个站点的子集具有同位雷达角反射器。大地测量网络的这个子集覆盖了 Dalby、Miles 和 Chinchilla 镇附近约 20,000 km 2 的区域。完整大地测量网络的永久安装已于 2014 年 11 月完成。计划每年对测量标记以及测量标记与角反射器之间的局部联系进行为期 7 天的占领活动 GNSS 测量。图 1. 昆士兰州南部 65 个新大地测量点的分布。图中还标出了指定为石油租赁的地块，用于石油、煤层气井和其他石油井的生产和勘探许可证（截至 2014 年 12 月 11 日）。 3. 角反射器设计 我们寻求一种单一目标设计，该设计能够在 XC 和 L 波段 SAR 图像数据中提供明亮而稳定的响应，并且在长时间（通常是十年）内保持稳定，以便在景观中永久部署。在昆士兰偏远地区部署时，选择了无源角反射器而不是有源应答器。一旦部署，角反射器就相对独立，因为转发器需要定期现场访问以进行长期维护、供电连续性和传输许可。由于合适的应答器[例如7] 目前尚未商用，设计 _____________________________________ Proc. “Fringe 2015 Workshop”，意大利弗拉斯卡蒂，2015 年 3 月 23-27 日（ESA SP-731，2015 年 5 月），角反射器的制造成本也低得多。此外，角反射器可以更轻松地满足我们对不同雷达频率的使用要求。选择三角形三面角反射器设计是因为制造简单、长期结构刚性、大雷达横截面的相对稳定性以及大约 40 度的三分贝波束宽度。最后一个事实确保永久部署的角反射器可以针对所有感兴趣的轨道 SAR 传感器（在上升或下降轨道通过）以平均视轴方向对齐，但每个 SAR 传感器的雷达响应的信噪比仍然很高且稳定。由于雷达响应取决于目标尺寸和雷达频率，我们认识到单一设计需要针对某些雷达频段进行折衷。我们进行了原型设计工作，以确定三角形三面角反射器的最合适尺寸[8]。我们制造了 18 个原型，内腿尺寸从 1.0 m 到 2.5 m 不等。一定视角范围内的雷达截面（RCS）是在地面雷达反射范围内表征的。在此表征之后，原型于 2014 年初部署在堪培拉附近的临时网络中。在此部署期间，使用来自 TerraSAR-X、COSMO-SkyMed、RADARSAT-2 和 RISAT-1 的 X 和 C 波段图像测试了原型的雷达响应。不幸的是，临时部署早于新的 L 波段 SAR 卫星 ALOS-2 的发射。这些测试的结果导致为永久苏拉特盆地大地测量网络选择了 1.5 m 内腿长度的角反射器（图 2）[8]。三个 2.0 m 和三个 2.5 m 原型角反射器被永久重新部署在苏拉特盆地大地测量网络中。永久部署的角反射器的第二个目的是为国际空间界提供一种可靠的方法来执行持续的辐射测量、几何测量和脉冲响应测量，以校准星载或机载平台上的 SAR 传感器[9]。迄今为止，ISRO、e-GEOS、JAXA 和 ESA 已使用它们来校准和/或验证 RISAT-1、COSMO-SkyMed、ALOS-2 和 Sentinel-1A SAR 数据。在苏拉特盆地大地测量网络中，大多数角反射器安装在沉积盆地的土壤剖面上；四十个角反射器之一与地面的基岩相连。每个角反射器的基础如图 2 所示，包括一块悬挂的 2 米见方的平板，由四根长度为 3 米的混凝土柱支撑。该地基旨在减轻上层土层季节性膨胀信号的影响。安装后，所有 40 个角反射器均已对齐，以便轨道 SAR 卫星的上升通过。计算每个位置的 ALOS2、RADARSAT-2 和 Sentinel-1A 的视轴方位角和仰角，并在对准每个角反射器时使用这些方向的平均值。 [6] 中给出了角反射器位置和当前对准。图2：a）1.5 m三角三面角反射器工程图（侧视图）和混凝土基础设计。尺寸以毫米为单位。 b) 在站点 26 安装 1.5 m 角反射器。混凝土板尺寸为 2.0 m 见方。大多数场地都有塑料围栏，以保护角反射器免受牲畜侵害
The combination of continuous Global Navigation Satellite System (GNSS) measurements over a sparse network of points covering Australia with relatively low frequency but high spatial density observations from Interferometric Synthetic Aperture Radar (InSAR) is fundamental to the new geodetic reference frame being developed for Australia. Recognising the economic importance of improved positional accuracy and the potential for geodetic tools to contribute to an understanding of energy related issues, the Australian Government has funded an innovative regional geodetic network of GNSS survey marks and co-located radar corner reflectors. This new network has been installed in the Surat Basin, Queensland where regional subsidence is expected due to significant resource extraction from the subsurface. In this contribution we present initial observations of the a-priori line-of-sight height error derived from corner reflector response in TerraSAR-X, Sentinel-1A, RADARSAT-2 and ALOS-2 SAR imagery of the Surat Basin. 1. COMBINATION OF GNSS AND INSAR Permanently deployed corner reflectors allow precise ground movements to be extracted from SAR acquisitions as persistent scatterers with stable phase characteristics at known positions [1, 2]. Since the signal to noise ratio from a resolution cell containing a corner reflector will be much greater than that from a resolution cell containing natural scatterers, the persistent scatterer signals can be used to validate the deformation signal measured spatially in SAR scenes using conventional differential InSAR time series techniques [3, 4]. The co-location of corner reflectors with GNSS survey marks will ensure SAR-derived deformation products can be tied to (and help constrain via deformation models) the local geodetic reference frame. 2. SURAT BASIN GEODETIC NETWORK The increasing demand for energy in Australia has led to increased exploitation of unconventional coal seam gas reserves, particularly in the Surat and Galilee basins in Queensland. CSG production began in the Surat Basin in 2006 and reserves are currently being produced by several operators. Predictions of the magnitude of subsidence based on poroelastic modelling and gas production rates indicate subsidence on the order of a decimetre may be occurring [5]. Due to this expected signal of around 100 km wavelength, we have deployed a new permanent geodetic network over the northern Surat Basin in Queensland (Fig. 1) [6]. The network consists of 65 GNSS survey mark sites, with a subset of 40 sites having a co-located radar corner reflector. This subset of the geodetic network covers a region of approximately 20,000 km 2 in the vicinity of the towns of Dalby, Miles and Chinchilla. Permanent installation of the full geodetic network was completed in November 2014. Annual 7-day occupation campaign GNSS measurements on the survey marks and local ties between survey marks and corner reflectors are planned. Figure 1. Distribution of 65 new geodetic sites in southern Queensland. Also plotted are land parcels designated as petroleum leases for production and exploration permits for petroleum, coal seam gas wells and other petroleum wells (as of 11 Dec 2014). 3. CORNER REFLECTOR DESIGN We sought a single target design that would provide a bright, yet stable response in XCand L-band SAR imagery data and that would be stable over long (nominally decadal) time periods for permanent deployment in the landscape. A passive corner reflector was chosen over an active transponder for deployment in remote Queensland. Once deployed, corner reflectors are comparatively autonomous since transponders would require regular site visits for maintenance, continuity of power supply and transmission licencing over long time periods. Since suitable transponders [e.g. 7] are not currently commercially available, the design _____________________________________ Proc. ‘Fringe 2015 Workshop’, Frascati, Italy 23–27 March 2015 (ESA SP-731, May 2015) and manufacture costs are also much less for a corner reflector. Furthermore, a corner reflector can more easily meet our requirement for use with different radar frequencies. A triangular trihedral corner reflector design was chosen because of the simplicity of manufacture, long term structural rigidity, relative stability for large radar cross section and a three decibel beam width of approximately 40 degrees. This last fact ensures that the permanently deployed corner reflectors can be aligned in an average boresight orientation for all orbiting SAR sensors of interest (on either an ascending or descending orbital pass) yet the signal to noise ratio of the radar response for each SAR sensor will still be high and stable. Since radar response is target size and radar frequency dependent we recognised that a single design would require a compromise for some radar frequency bands. We undertook a prototyping exercise to establish the most appropriate size of triangular trihedral corner reflector [8]. We manufactured 18 prototypes ranging in size from 1.0 m to 2.5 m inner leg dimension. The radar cross section (RCS) at a range of viewing angles was characterised at a ground radar reflection range. Following this characterisation, the prototypes were deployed in a temporary network near to Canberra in early 2014. During this deployment the radar response of the prototypes was tested with Xand C-band imagery from TerraSAR-X, COSMO-SkyMed, RADARSAT-2 and RISAT-1. Unfortunately the temporary deployment pre-dated the launch of the new L-band SAR satellite, ALOS-2. Results from these tests led to a corner reflector of 1.5 m inner leg length being chosen for the permanent Surat Basin geodetic network (Fig. 2) [8]. The three 2.0 m and three 2.5 m prototype corner reflectors were permanently re-deployed in the Surat Basin geodetic network. A secondary aim of the permanently deployed corner reflectors is that they provide the international space community with a reliable means to perform ongoing radiometric, geometric, and impulse response measurements for calibration of SAR sensors on spaceborne or airborne platforms [9]. To date they have been used by ISRO, e-GEOS, JAXA and ESA to calibrate and/or validate RISAT-1, COSMO-SkyMed, ALOS-2 and Sentinel-1A SAR data. In the Surat Basin geodetic network, most corner reflectors are installed on the soil profile of the sedimentary basin; one of the forty corner reflectors was coupled to bedrock at the ground surface. The foundation for each corner reflector, depicted in Fig. 2, comprises a suspended two metre square slab supported by four concrete pillars each of 3.0 metre length. This foundation was designed to mitigate the impact of signals originating from seasonal swelling in the upper soil layers. Upon installation, all forty corner reflectors were aligned for ascending passes of orbiting SAR satellites. The boresight azimuth and elevation angles for ALOS2, RADARSAT-2 and Sentinel-1A were calculated for each position, and an average of these orientations was used when aligning each corner reflector. Corner reflector positions and current alignments are given in [6]. Figure 2: a) Engineering drawing of the 1.5 m triangular trihedral corner reflector (side-on view) and concrete foundation design. Dimensions in mm. b) 1.5 m corner reflector installed at Site 26. Concrete slab dimension is 2.0 m square. Most sites have a plastic fence to protect the corner reflector from livestock