最大熵谱估计中振幅谱和相位谱的求法

EXTRACTION OF AMPLITUDE AND PHASE SPECTRUM BY MAXIMUM ENTROPY ESTIMATION

也质矿产部北京计算中心

最大熵谱估计克服了窗函数对传统谱估计的不利影响,特别在数据短的情况下,它比传统谱估计优越得多。但最大熵谱估计所确定的信号最佳频率成分,其对应的振幅值并不反映原来振幅值,且随自回归模型的阶数的不同而变化。这是最大熵谱估计法的最大缺点,也是此法目前还不能很好地用于地震数据处理的主要原因。本文介绍两种由得到的最佳频率成分来恢复原振幅值和相位值的方法。

The harmful effects of the window function on the conventional spec-tram estimation could be eliminated by the use of maximum entropy estimation. The technique recommended has many advantages over the conventional ones, especially in the case of short data. However, the amplitudes corresponding to the optimum signal, frequencies determined by the maximum entropy estimation certainly do not reflect the original ones, in fact, they vary with the exponent of autoregressin model. This is the main deficiency of the technique-the maximun entropy estimation and the major reason that we couldn't still makd good use of it in seismic data processing. Two techniques for the recovery of original amplitudes and phases through the optimum frequencies obtained were introduced in this paper.