基于InSAR技术的树坪滑坡识别与研究

发布时间：2018-11-20 12:06

【摘要】：三峡库区地质条件复杂,同时年降雨量多在1000mm以上,多大雨、暴雨等强降水过程。古滑坡数量大,且自2003年三峡大坝蓄水以来新的滑坡体也在不断增加,成为三峡库区最为严重的地质灾害之一。由于三峡库区滑坡的突发性和隐蔽性等特点,常规手段很难发现和监测,给生态和人类的生存、生活带来了很大的危险和隐患。雷达遥感技术由于其全天时、全天候、大范围以及不受云雨等气候影响,因此作为库区滑坡监测的新技术、新手段,InSAR技术越来越被重视和应用。然而,由于三峡库区地形起伏比较大,植被繁多,在实际的数据处理应用中,受时间去相干、空间去相干及大气扰动等影响明显,给滑坡的形变提取带来很大的困惑。另外D-InSAR技术只能提取视线向上的形变信息,不能直观反映真实的三维形变特征。针对这些问题,本文展开了系列的分析和探讨,主要研究内容和贡献包括:(1)本文利用干涉图堆叠技术(Stacking)对三峡库区秭归地区大范围区域滑坡进行了探测与识别,获取了年平均形变速率在15cm以上的缓慢形变区域共6处,其中5处为已知滑坡,1处为疑似滑坡。(2)利用SBAS技术获取了树坪滑坡2007年至2011年间的时间序列形变信息,揭示了树坪滑坡东西两个滑坡体的形变特征,并与水库水位的变化进行了相关性分析,为进一步揭示形变机理提供了参考。(3)结合升降轨数据,利用像素偏移量技术获取方位向和距离向的二维形变信息。在此基础上,分别采用等精度、不等精度和基于方差分量估计的三维形变信息提取。结果表明,基于方差分量估计的三维形变信息能很好地调整数据源的权比,拟合精度得到了较大的改善。(4)考虑到不同数据源间存有系统性误差,尝试附加系统参数的三维形变信息提取,拟合精度也得到了相应的改善。
[Abstract]:The geological conditions in the three Gorges reservoir area are complicated, and the annual rainfall is more than 1000mm, heavy rain and heavy rain. The number of ancient landslides is large and the new landslide body has been increasing since the three Gorges Dam was impounded in 2003. It has become one of the most serious geological disasters in the three Gorges Reservoir area. Because of the sudden occurrence and concealment of landslide in the three Gorges Reservoir area, it is difficult to detect and monitor the landslide by conventional means, which brings great danger and hidden danger to the survival of ecology and human beings. Radar remote sensing technology is more and more important and applied as a new technique and means of landslide monitoring in reservoir area, because of its weather, wide range and no influence by cloud and rain. However, because of the large topographic fluctuation and various vegetation in the three Gorges Reservoir area, in the practical data processing application, it is obviously affected by time decoherence, spatial decoherence and atmospheric disturbance, which brings great confusion to the deformation extraction of landslide. In addition, the D-InSAR technique can only extract the deformation information of the line of sight upward, and can not reflect the true 3D deformation characteristics directly. Aiming at these problems, a series of analysis and discussion are carried out in this paper. The main research contents and contributions are as follows: (1) in this paper, the interference map stacking technique (Stacking) is used to detect and identify the landslide in a large area in Zigui area of the three Gorges Reservoir area. A total of 6 slow deformation regions with annual average deformation rate above 15cm are obtained, of which 5 are known landslides and 1 is suspected landslide. (2) time series deformation information of Shuping landslide from 2007 to 2011 is obtained by using SBAS technique. The deformation characteristics of the two landslide bodies in the east and west of Shuping landslide are revealed, and the correlation analysis with the change of reservoir water level is carried out, which provides a reference for further revealing the deformation mechanism. (3) combined with the data of ascending and descending rail, The two-dimensional deformation information of azimuth and distance is obtained by pixel offset technique. On this basis, three dimensional deformation information is extracted with equal precision, unequal precision and variance component estimation. The results show that the 3D deformation information based on variance component estimation can adjust the weight ratio of the data source very well, and the fitting accuracy is greatly improved. (4) considering the systematic error between different data sources, Three-dimensional deformation information extraction with system parameters is attempted and the fitting accuracy is improved accordingly.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P642.22

【相似文献】