子像素偏移追踪算法及其在实际地震同震形变场监测中的应用

发布时间：2018-05-09 02:39

本文选题：遥感影像 + 子像素偏移追踪　；参考：《中国科学技术大学》2015年硕士论文

【摘要】：地震是危害人类生产、生活的重大灾害现象,对于地震的研究是很有必要的。大地震地表破裂的观测资料对于研究地震力学和断层区域结构性质都具有重要意义。断层几何结构和相关滑移分布也是决定地震动力学破裂过程的关键所在。利用地表形变位移场可以反演断层的滑移分布,研究断层结构和断层破裂状态；同时也可以将地表位移场作为边界条件,对地震断裂运动进行约束。因此对于研究地震动力学和运动学问题来说,获取大地震同震位移场都是至关重要的科学问题。本文在现有影像偏移追踪算法的基础上,结合强度互相关、双线性插值和二维高斯回归三种算法,提出了一种子像素影像偏移追踪算法对地震同震位移场进行监测。我们将该算法利用C语言程序化,并将该算法应用到实际地震中去,求取了2014年2月14日新疆于田地震和2013年9月24日巴基斯坦地震的同震形变偏移场,并对所得位移场进行掩膜、滤波、差分等后处理,得到准确平滑的同震偏移场。2014年于田地震震中东侧有两个条带存在左旋走滑趋势的断层,分别为北硝尔库勒断裂和阿什库勒-硝尔库勒断裂,断裂以左旋走滑为主,并带有一定量的正滑分量,两条断裂长约20km,整体走向约呈N60。E,最大偏移量可达到约1.5m,与其他研究者得到结果相一致；在震中南侧,余震主要分布区域也存在一条近南北向的弧形断裂。2013年巴基斯坦地震地表破裂主要表现为左旋走滑的弧形破裂,平均滑移量6米,最大可达到10米,断层西北侧的滑移量明显大于东南侧的滑移量。并且给出了在该逆断层发生走滑型地震的可能原因。本文得到的巴基斯坦地震同震形变位移场与前人的研究结果基本一致,精度可达到百分之三个像素水平。两个实际地震中的应用结果表明新发展的子像素影像偏移追踪算法可以有效求取中强震的同震形变位移场。
[Abstract]:Earthquake is a major disaster phenomenon that endangers human production and life, so it is necessary to study earthquake. The observational data of surface rupture of large earthquakes are of great significance for studying seismic mechanics and regional structural properties of faults. The geometrical structure of faults and the relative slip distribution are also the key factors to determine the process of seismic dynamic rupture. The surface deformation displacement field can be used to invert the slip distribution of the fault and study the fault structure and fracture state, and the surface displacement field can be taken as the boundary condition to constrain the seismic fault movement. Therefore, it is very important to obtain the coseismic displacement field of large earthquakes for the study of seismic dynamics and kinematics. Based on the existing image migration tracking algorithms, combined with intensity cross-correlation, bilinear interpolation and two-dimensional Gao Si regression, a subpixel image migration tracking algorithm is proposed to monitor the seismic coseismic displacement field. We use C language to program the algorithm, and apply the algorithm to the actual earthquake, and obtain the coseismic deformation migration field of the Yutian earthquake in Xinjiang on February 14, 2014 and Pakistan earthquake on September 24, 2013. The obtained displacement field is processed by mask, filtering, difference and so on, and an accurate and smooth coearthquake migration field is obtained. There are two faults with left-lateral strike-slip trend in the east side of the 2014 Yutian earthquake epicenter. The faults are mainly left-lateral strike-slip faults with a certain amount of normal slip components, respectively, for the North Nielkul fault and the Ashkul-Nierkul fault, which are mainly left-lateral strike-slip faults. The length of the two faults is about 20 km, the whole strike is N60. 0. E, the maximum migration is about 1. 5 m, which is consistent with the results obtained by other researchers, in the south side of the epicenter, There is also an arc fault in the main distribution area of aftershocks. The surface rupture of the 2013 Pakistan earthquake mainly shows the arc rupture of left-lateral strike-slip, with an average slip of 6 meters and a maximum of 10 meters. The slip amount on the north and west side of the fault is obviously larger than that on the southeast side. The possible causes of strike-slip earthquakes on the thrust fault are also given. The displacement field of the earthquake coearthquake deformation obtained in this paper is basically consistent with the previous research results, and the accuracy can reach 3 percent. The application results of two practical earthquakes show that the newly developed subpixel image migration tracking algorithm can effectively obtain the coseismic deformation displacement field of moderate and strong earthquakes.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P315.7

【参考文献】