带乘性噪声系统的最优反褶积

Optimal Deconvolution for Systems with Multiplicative Noise

针对高分辨率石油地震勘探问题，研究一种带乘性噪声随机系统的最优反褶积估计方法。通过建立逆向运行的在线性最小方差意义下最优的状态滤波公式，推导出一种随机输入序列的单向最优固定域递推算法。该方法离线运算的存储最小，有助于提高地层反射序列提取的分辨率。

In the light of seismic oil exploration problems with high resolution, an optimal deconvolution method for systems with multiplicative noise is presented in the sense of linear minimumvariance. We develop the deconvolution algorithm in two steps. First, an optimal fixedinterval smoother of reversetime is constructed based on the reversedtime state space model for the systems. Then a onepass optimal deconvolution algorithm is presented by means of the state smoother. This deconvolution algorithm can avoid high storage requirements, and the simulation results for seismic signal processing are satisfactory.