项目名称: 稀疏约束方法在探地雷达反演与成像中的应用
项目编号: No.41304093
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 天文学、地球科学
项目作者: 丁亮
作者单位: 东北林业大学
项目金额: 25万元
中文摘要: 目前,探地雷达数据解释工作仍处于凭借经验对探测结果进行人工判读的水平,虽然稀疏约束方法在探地雷达反问题中已经得到了广泛的应用,但如何提高反演成像的分辨率仍然是核心问题。传统的小波稀疏工具只能逐点检测物体内部反演参数,且只能对点状的内部结构具有较好的反演结果。剪切波作为新兴的多方向、多尺度稀疏工具可以准确、高效地检测地球内部不规则的层状边界,由于复杂的地质结构往往包含点状及层状介质,我们针对探地雷达资料反演与成像问题,以Maxwell方程为研究对象,开展基于剪切波和小波稀疏约束的反演理论研究。构造稀疏约束正则化泛函,给出两种约束的权重选取方法,研究反演参数的稀疏性表示,从而极大地减少计算量和对数据的需求,在识别地球内部各介质边界的同时,有效地识别小的异常体,例如尖、角和振荡边界,从而实现复杂介质的高分辨率反演成像。本项目的研究方法是全新的,对于其他复杂介质反演问题也具有适用性。
中文关键词: 探地雷达;反问题;正演模拟;稀疏约束;信号处理
英文摘要: At present, the ground penetrating radar (Ground Penetrating Radar) data interpretation technique is still established on the experience for radar image identification and manual delineated radar abnormal body. Although sparsity constraint method has been widely applied to GPR inverse problems, how to improve the resolution is still the core problem. Tranditional wavelet sparsity can only detect the inversion parameter pointwise, and have effective inversion result only when interior structure is punctiform. Shearlet is a emerging sparsity fool which has multi-direction and multi-scale, it can detect interior irregular layer boundary effectively. Complicated geologic structure always includes layer and punctiform medium. According to the inversion and imaging problem of GPR, we study the Maxwell equation and research the sparsity constraint inversion based on shearlet and wavelet, construct sparsity constraint regularization functional, give the weight selection method for two constraints. In order to reduce the calculated quantities and the demand of the data, we research sparsity presentation of inversion parameter. This structure can distinguish the boundary of mediums, at the same time distinguish anomalous structure for example point, corner and oscillating boundary to achieve high inversion resolution for
英文关键词: ground penetrating radar;;inverse probelms;forward simulation;sparsity constraint;signal processing