联合永久散射体雷达干涉和像素偏移估计方法提取火山形变及参数反演

发布时间：2018-03-08 00:01

本文选题：永久散射体雷达干涉　切入点：像素偏移估计　出处：《西南交通大学》2015年硕士论文　论文类型：学位论文

【摘要】：火山灾害是最严重的自然灾害之一,通常伴随着一系列的地表形变,这些形变在一定程度上能够反映火山运动的特征以及火山内部压力源的扰动情况。因此,探测此类形变可为监测火山运动提供依据,对预防或减少火山运动带来的危害和损失具有重要意义。目前,在众多地表形变探测技术中,因基于SAR影像分析的微波遥感技术具有高精度和大范围测量等优势,已被广泛地应用于火山形变的探测。但随着研究的深入发现,该技术在火山形变探测过程中,因火山地区裸地多、形变梯度大,受影像配准精度和形变探测模型等因素的影响,导致失相干严重、形变探测精度低,进而无法为火山参数反演提供可靠的形变信息。为了给火山参数的正确反演提供可靠的地表形变信息,本论文以谢拉·内格拉火山为实验区,在已有的像素偏移估计方法和PSI技术的基础上,从提高形变探测精度和空间覆盖度两方面着手,提出了组合模型和干涉对组合的优化方法,在火山地区获取了高精度和空间覆盖度的地表形变信息。论文具体工作如下：(1)本论文以精确配准后的影像可获得更可靠的形变偏移量为出发点,以配准精度与空间基线具有相关性为依据,引入最小二乘法参与配准以提高时序SAR影像的配准精度,为像素偏移估计技术提供了高精度的配准数据源；(2)基于火山形变演变特征,提出以组合模型作为形变解算的约束条件,并将其应用在时序像素偏移估计。通过GPS数据的交叉验证表明,提出的组合模型切实有效,并且很好地反映了地表水平向形变的演变趋势,其在方位向和距离向的精度可达0.070 m和0.160 m,相较于传统时序模型的探测精度有大幅提升。(3)在PSI技术探测火山形变过程中,SAR影像的干涉组合方式十分重要,直接影响到干涉质量与测量精度。为了获取火山口地区的有效形变信息,论文在常规干涉组合方式的基础上进行干涉组合优化,获取了该火山口的地表形变场。同期GPS数据验证结果表明,最终精度提升至厘米级。(4)采用上述实验区中不同方向的地表形变信息为约束条,引入Mogi模型反演得到火山压力源参数,并针对不同方向的形变以及组合模型讨论了反演火山岩浆房深度与岩浆等效体积变化的影响。其中,以组合模型获取的水平形变进行反演显示,在火山喷发前等效体积变化速率为1.59×107 m3/year,在火山喷发期间最大等效体积变化约为1.02×108 m3。并且,该模型受噪声影响较小,以其形变结果反演得到的岩浆体积变化趋势与传统时序模型相比,波动比较平缓。最终反演结果表明,联合两种方法与组合模型,可为火山活动参数的反演提供更有效、全面的地表约束条件,其反演结果更具有可靠性。通过理论与实验研究表明,本文提出的组合模型与干涉组合优化方法提高了火山地表形变的探测精度和空间覆盖度,为火山参数的正确反演提供可靠的地表形变信息。
[Abstract]:The volcano disaster is one of the most serious natural disasters, usually accompanied by a series of surface deformation, deformation of these disturbances can reflect the characteristics of volcano movement and volcano internal pressure source to a certain extent. Therefore, detection of such deformation can provide the basis for monitoring the volcano movement, has important significance to prevent or reduce the volcano movement harm and loss. At present, in many surface deformation detection technology, the microwave remote sensing technique based on SAR image analysis with high accuracy and large measurement range and other advantages, has been widely applied to the detection of shape change. But the volcano with in-depth study found that the technology in the process of detecting deformation due to volcano, volcano area the bare land, the deformation gradient, affected by the image registration precision and deformation detecting model and other factors, resulting in serious loss of coherence, deformation detection precision is low, and no method for parameters of volcano The inversion provides reliable information. The deformation of surface deformation provide reliable information in order to correct the inversion parameters to the volcano, the volcano Sheila Negra into the experimentation area, based on the pixel offset estimation method and PSI technology, in order to improve the detection accuracy and spatial coverage of deformation in two aspects, put forward the combination model and interference optimization method in combination, in the volcano area to obtain surface deformation information of high accuracy and spatial coverage. The specific work is as follows: (1) the accurate registration of the image after deformation can be offset more reliable as the starting point, with the registration accuracy and spatial baseline correlation based on the least square method in registration in order to improve the timing of SAR image registration accuracy for pixel offset estimation technology provides high precision registration data source; (2) based on the characteristics of the evolution of the volcano deformation. With the combined model as constraint deformation calculation, and its application in the estimation of timing. The pixel offset by cross validation of GPS data show that the combination model proposed is effective, and can well reflect the evolution trend of surface horizontal deformation, the range accuracy can reach to 0.070 and the distance to m and 0.160 m, compared to the detection accuracy of the traditional time series models have increased dramatically. (3) in the process of detecting volcano deformation with PSI, a combination of SAR interference image is very important, directly affects the quality of interference and precision. In order to obtain deformation information of crater area, the interference of combinatorial optimization based on conventional interference the combination, the surface deformation field of the volcano. The same period GPS data show that the final accuracy to centimeter level. (4) the surface of different parties in the experimentation area to the above Information as constraints, obtained volcano pressure source parameters into Mogi model inversion, and discuss the effect of inversion depth and volcano magma chamber volume changes in magma equivalent deformation in different direction and the combination model. Among them, the horizontal deformation is obtained by model inversion shows that the eruption of the volcano before the equivalent volume change rate is 1.59 * 107 m3/year, the eruption of the volcano during the maximum equivalent volume change is about 1.02 * 108 and m3., the model is less affected by noise, its deformation results obtained by inversion of the magma volume change trend compared with the traditional time series model, volatility is relatively slow. The inversion results show that combined with the combination of two kinds of methods for model inversion volcano activity parameters provide more effective, comprehensive ground constraints, the inversion result is more reliable. The theoretical and experimental results show that the combination of the The combined optimization method of model and interference improves the accuracy and coverage of volcanic surface deformation, providing reliable surface deformation information for correct inversion of volcanic parameters.

【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P317.6

【相似文献】