Shape detection of multiple subsurface cavities by particle filtering with elastic wave propagation
Shape detection of multiple subsurface cavities by particle filtering with elastic wave propagation
复制标题
DOI：
10.1002/nag.3117
复制
发表时间：
2020-07
影响因子：
4
通讯作者：
R. Takamatsu;K. Fujisawa;K. Nakahata;A. Murakami
中科院分区：
文献类型：
--
作者：
R. Takamatsu;K. Fujisawa;K. Nakahata;A. Murakami
将粒子滤波和参数水平集方法应用于地下弹性波传播，提出了一种检测地下孔洞数量和形状的数值方法。当地下孔洞存在时，在地下传播的弹性波在孔洞的边界面处反射。通过观测地表反射波的传播速度，并利用参数化水平函数对空洞的形状进行参数化，粒子滤波可以识别空洞的数量和形状。利用合成观测资料进行了多空洞探测的数值试验。结果表明，该方法可以通过后验概率分布函数中的峰数来识别空洞的个数，并通过水平集函数的参数识别来估计空洞的形状，从而解决几何反问题.
A numerical technique for detecting the number and shape of subsurface cavities is presented, applying the particle filter and the parametric level set method to elastic wave propagation under the ground. When subsurface cavities exist, the elastic wave propagating in the ground is reflected at the boundary faces of the cavities. Observing the velocity of the reflection wave at the surface of a ground that includes multiple cavities and parameterizing the shape of the cavities by the parametric level function, both the number and the shape of the cavities can be identified by the particle filter. Numerical experimentation for detecting multiple cavities is conducted with synthetic observation data. The results show that the proposed technique enables the number of cavities to be identified by the number of peaks in the posterior probabilistic distribution function and solves geometric inverse problems by estimating the shape of the cavities through the parameter identification of the level set function.