基于褶积模型的多尺度联合反演问题

Multiscaled-combined inversion problem based on convolution model

本文基于褶积模型推导出多尺度约束最小二乘解，提出小波域多尺度联合反演方法(MCIM, MultiscaleCombined Inversion Method)。利用多尺度间的内在联系，采用尺度序列分 解，以粗尺度下的反演解作为待优化的当前尺度反演的约束，提高反演效率与稳定性。多组 测试数据显示，对于薄层检测与去噪，该方法具有良好的效果。

Based on convolution model,the paper deduced least square solution with multi scale constraint and proposed a multiscale combined inversion method in wave let domain.Using internal connection between multiscale,adapting decomposit ion of scale series and putting solution of inversion under sparse scale as a co nstraint of waiting for optimal and current scale inversion can improve effect iveness and stability of inversion.Multi group test data showed that it has ad vantages of good thin layer detection,etc.