白噪声Wiener反卷积滤波器

White Noise Wiener Deconvolution Filters

应用现代时间序列分析方法 ,基于 ARMA新息模型提出了白噪声 Wiener反卷积滤波器。该滤波器可统一处理滤波、平滑和预报问题 ,可用 ARMA递推滤波器实现 ,适用于石油地震勘探数据处理。同多项式方法和 Kalman滤波方法相比 ,避免了求解 Diophantine方程和 Riccati方程 ,减少了计算负担。Bernoulli- Gaussian白噪声反卷积的仿真例子说明了其有效性。

Based on the autoregressive moving average (ARMA) innovation model, the white noise Wiener deconvolution filters are pressented by using the modern time series analysis method. The white noise filtering, smoothing and prediction problems can be dealt with in a unified framework. They can be implemented via the ARMA recursive filters and can be applied to data processing in oil seismic exploration. Compared with the polynomial method and Kalman filtering method, they avoid solving the Diophantine equations and Riccati equations, so that the computational burden is reduced. A simulation example for Bernoulli Gaussian white noise deconvolution shows their effectiveness.