一种适应于海量地震数据时频分析的快速算法

摘要

为了克服MATLAB时频工具箱在计算大容量的地震数据的分频处理时,存在计算速度慢、内存不足以及FFT的计算长度严重影响时频分辨率等问题,使其适应于实际海量地震数据的时频分析.对MATLAB时频工具箱中基于离散短时傅里叶变换的时频算法进行了改进,采用Fortran语言在微机上编程实现了该算法.理论实算表明,改进后的Fortran程序能够高效地计算大容量、短时窗地震数据的分频处理.%As issues of low calculation speed, out of memory, and the calculation length of FFT seriously affecting the capacity of time-frequency resolution, etc. , happened during calculating of frequency decomposition on large-capacity seismic data. In order to overcome those issues, this paper improved MATLAB time-frequency toolbox' s time-frequency algorithm, which based on the discrete short time Fourier transform, implemented the algorithm by applying of Fortran language in microcomputer programming. The theory calculation tests show that the improved Fortran program can efficiently calculate the frequency decomposition of a large-capacity, short-term window seismic data.