Developing observation techniques with high accuracy to monitor weak changes in medium is significant important to understand evolution of the Earth. Recently, a new coda wave differential imaging technique named Locadiff has the ability to detect local changes of the medium with velocity changes and scatterers change based on coda wave propagation model under multiple scattering theory. The method avoid the limitation of coda wave interferometry which can only provide globe change information of the medium.. With priori information constraint and maximum likelihood matching between observation data and forward model by likelihood function, the Bayesian inversion technique is capable to avoid disadvantages in resolution, accuracy and results evaluation over existing linear least-squares method. . In this proposal, 4 tasks are carried out to improve the accuracy and resolution of coda wave differential imaging technique including (1) implementation on coda wave differential imaging technique based on Bayesian inversion; (2)a comparison study on resolution,accuracy and performance on inversion results between exist method and Bayesian base method at different noise levels with numerical simulation and ultrasound experiment on rocks; (3)study on changes with different geometry. Obtain and analysis the inversion results on changes with different length and orientations. Find out the influence on inversion as well as to propose the improvement; (4) discussion on it's possible application to monitoring changes in regional scale of the Earth.
研究介质动态变化需要发展高精度观测技术。新近的尾波差分成像方法提出了描述台站间尾波在介质中的传播模型,能够对介质内局部动态变化成像。该方法摆脱了尾波干涉法只能观测介质整体动态变化的局限。贝叶斯反演技术通过先验信息约束和似然函数匹配反演模型参数在现有观测数据下最大可能分布,有效避免尾波差分成像方法中现有反演技术在精度、可靠性及反演结果评价等方面的缺陷。因此本项目拟开展:(1)实现基于贝叶斯反演的尾波差分成像方法。(2)结合数值模拟及岩石超声观测实验开展与现有反演技术对比研究。获得不同观测噪声下反演结果。对比精度和误差的改善程度。(3)开展具有几何特征的动态变化观测研究。获得不同长度、角度扩展型变化的反演结果。分析对反演结果的影响并改进。(4)讨论利用该方法开展区域尺度地下介质动态变化成像研究的可行性方案。以上研究将为利用该方法开展区域尺度地学问题研究提供基础。
本课题按规定完成所有计划的研究任务和目标。形成的基于贝叶斯反演技术的尾波差分成像方法可将动态微裂隙的成像精度提升到1/15波长,并且能以1e-6的观测精度对早期损伤介质开展健康检测。同时给出微裂隙变化几何特征的预测模型。为将来该方法的进一步发展提供基础。利用该方法在超声实验中能够以高时空分辨率追踪介质内部由应力变化导致的复杂的时空变化。这是以往传统方法无法达到的。这使得我们的研究在国际上处于领先地位,为下一步开展断层应力场动态变化监测以及滑坡灾害的监测和预警等方面的应用提供了坚实基础。本课题共产出论文5篇。其中英文SCI索引3篇,中文SCI索引1篇,英文EI索引一篇。
{{i.achievement_title}}
数据更新时间:2023-05-31
基于分形L系统的水稻根系建模方法研究
涡度相关技术及其在陆地生态系统通量研究中的应用
正交异性钢桥面板纵肋-面板疲劳开裂的CFRP加固研究
基于 Kronecker 压缩感知的宽带 MIMO 雷达高分辨三维成像
主控因素对异型头弹丸半侵彻金属靶深度的影响特性研究
基于贝叶斯稀疏约束的孔隙介质弹性波方程反演方法研究
基于贝叶斯神经网络的PDE参数反演
二类水体水色遥感反演的贝叶斯方法
基于稀疏贝叶斯方法的THz-SAR振动目标成像研究