InSAR技术形变监测中的干涉条件研究

发布时间：2018-12-26 15:03

【摘要】：自上世纪60年代以来,雷达干涉测量(InSAR)技术具有获取高精度空间三维信息的能力已经得到了证明,在其基础上发展而来的D-InSAR技术也逐渐成为了测量地形高程和地表形变的最有力手段之一。由于地表形变威胁着人类生存环境,对地表形变进行监测引起了各国政府的重视。InSAR/D-InSAR技术的出现,克服了人工水准测量、GPS技术和光学遥感等传统手段获取形变信息所受到的限制,实现了大面积、低成本、高效率地监测地面微小形变,监测精度可达厘米至毫米级。然而在D-InSAR处理过程中,该技术受到大气效应和失相关等因素的制约,大大影响了监测结果的精度和应用范围的推广。虽然随着D-InSAR技术的发展,在去除大气效应等领域取得了不少研究成果,但仍有一些因素制约了该技术的运用。因此,本文主要基于当前在D-InSAR约束条件分析方面的研究尚少的现状,深入地探讨影响D-InSAR形变监测的几个关键因素,结合模拟数据和真实数据论述D-InSAR技术的干涉条件,得到的结论将有助于建立更加完善的干涉限制数学模型和研究感兴趣区域时的数据选择及参数配置,并为未来研制新型SAR系统的干涉成像参数设计和干涉成像任务规划提供了科学依据。本文的研究内容主要分为以下几个方面： (1)量化分析了与SAR系统参数相关的几个分量(比如入射角、波长、影像分辨率、系统带宽)对实际观测的影响,并对不同SAR系统的形变监测能力进行了对比；明确给出了各个SAR系统的临界坡度范围及监测盲区；提出了新的形变梯度模型,引入了新参数入射角；证明了植被覆盖区,相干性分解需要考虑体散射分量；并对不同类型观测区适合的数据类型给出了具体的建议。 (2)验证了本文分析的干涉约束条件。通过大量模拟数据和真实数据,有针对性的进行检验。其中,用模拟数据验证了形变梯度、SAR几何成像特性和实际应用中所受到的各种约束限制条件,用三峡地区的ENVISAT ASAR和TerraSAR-X数据分析了相干性。 (3)提出了精确提取时间去相关分量的融合方法。即改进了自适应区域增长算法(IDAN)用于提取精确的相干系数；并利用最小二乘拟合去除估计量偏差,得到相干性的无偏估计；引入了相干性分解技术在最大程度上稳健地分解相干值,得到精确的时间去相关分量。 (4)验证了本文提出的提取时间去相关分量融合算法的有效性。理论上来说,新方法能够更好地反映出时间与时间去相关分量之间的函数关系,可得到稳健可靠相干性估计结果,因此本文通过ENVISAT ASAR数据集对新方法和常规方法进行对比实验,证明了实际应用中新的融合算法可以更好地对InSAR相干性进行估计。 (5)给出了一个根据D-InSAR干涉条件分析开展形变监测工作的实例。由于南极冰川对全球变化有着重要意义,冰流速提取是南极冰川监测的重要方向之一。而冰流场属于地表快速运动区域,针对这种类型的实验区,可采用L波段的数据进行监测。因此,本文选取了南极格罗夫山区的ALOS/PALSAR数据提取二维水平冰流速场。由于单轨SAR数据只能提供一个方向上的流速,因此我们采用D-InSAR技术探测地距向流速,Offset-tracking技术探测方位向流速,合成了该实验区的二维水平冰流速场,并与NASA/JPL发布的全南极冰流速和GPS实地测量结果进行交叉验证。该实验采用了D-InSAR干涉条件分析给出的数据选择建议,证明快速运动区域的形变监测使用L波段数据是非常有效的。
[Abstract]:Since the 1960s, the ability of the radar interferometry (InSAR) to obtain the three-dimensional information of high-precision space has been proved, and the D-InSAR technology developed on the basis of its development has become one of the most powerful means to measure the elevation of the terrain and the surface deformation. The surface deformation is a threat to the human living environment, and the monitoring of the surface deformation has caused the attention of the governments. The occurrence of InSAR/ D-InSAR technology has overcome the limitation of the traditional means of artificial leveling, GPS technology and optical remote sensing to obtain the deformation information, and the micro-deformation of the ground is monitored with large area, low cost and high efficiency, and the monitoring precision can reach the millimeter level. In the process of D-InSAR, however, the technology is restricted by the factors such as atmospheric effect and distortion, which greatly influences the accuracy and application range of the monitoring results. However, with the development of D-InSAR technology, many research achievements have been made in the field of removing atmospheric effect, but there are still some factors that restrict the transportation of the technology. In this paper, based on the current research of D-InSAR constraint analysis, the paper discusses several key factors that affect the D-InSAR deformation monitoring, and discusses the interference strip of D-InSAR technology in combination with the simulation data and the real data. The conclusions of this paper will help to establish a more perfect mathematical model of interference and study the data selection and parameter configuration in the region of interest, and provide the scientific basis for the design and interference imaging task planning of the new SAR system in the future. The research contents of this paper are mainly divided into the following: Surface: (1) The effect of several components (such as incident angle, wavelength, image resolution, system bandwidth) related to the parameters of the SAR system on the actual observation is analyzed and analyzed, and the deformation monitoring capability of different SAR systems is carried out. The critical slope range and the monitoring dead zone of each SAR system are given. The new deformation gradient model is put forward, and the incident angle of the new parameter is introduced. The vegetation coverage area and the coherence decomposition are considered to be considered. and the data types suitable for different types of observation areas are given in detail. The recommendations of this article are verified (2). Constraints. Targeted at a large number of analog data and real data The simulation data is used to verify the deformation gradient, the geometric imaging characteristics of the SAR and the various constraints imposed in the practical application, and the ENVISAT ASSAR and TerraSAR-X data are used in the Three Gorges area. The coherence is analyzed. (3) The accurate extraction time is proposed. The invention provides a fusion method of components, namely, an adaptive region growth algorithm (IDAN) is improved for extracting an accurate coherence coefficient, and the estimation deviation is removed by a least two-by-one fit, so that an unbiased estimation of the coherence is obtained; and the coherence decomposition technology is introduced to a maximum degree, robust decomposition of the coherent values to obtain an accurate Time to correlation component. (4) The extraction time proposed in this paper is verified to be related. In theory, the new method can better reflect the function relation between time and time decorrelation component, and can get the result of robust and reliable coherence estimation. It is proved that the new fusion algorithm in the practical application can be better applied to the InS. AR coherence is estimated. (5) An analysis of the interference conditions based on D-InSAR is given. An example of deformation monitoring is carried out. Because the Antarctic ice is of great significance to global change, the extraction of ice flow rate is the antarctic The ice flow field is one of the important direction of the glacier monitoring, and the ice flow field belongs to the surface fast moving area, which can be used for the experimental area of this type. The data of L-band is used for monitoring. Therefore, the number of ALOS/ PALSAR in the Antarctic Grove Mountain area is selected. The two-dimensional horizontal ice flow field is extracted. Since the single-rail SAR data can only provide a flow rate in one direction, we use the D-InSAR technology to detect the directional flow velocity, the offset-tracking technique, and to synthesize the real-time flow velocity. Two-dimensional horizontal ice flow field for the test area and the full Antarctic ice flow rate and GPS real-time with NASA/ JPL The results are cross-validated. The results of the data selection based on the D-InSAR interferometry are used to demonstrate the use of the deformation monitoring in the fast moving area.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P225.1

【参考文献】