地震数据时频分析与分频处理

在地震数据处理过程中，时频分析技术可较好地描述信号和噪声的频率随时间变化的规律，即时频分析可以细致地刻画数据中各类波的时频结构和性质，其重要性在于能直观地识别、检测和分类数据，因此，可采取有针对性的措施，以提高每一个步骤的处理精度。简述了时频分析的方法原理，介绍了适应于地震数据分析的窗函数的选取方法，并展示了分频噪声压制的应用效果，说明分频处理是提高数据处理保真度的有效方法。

Time-frequency analysis and frequency-divisional processing of seismic data

Time-frequency analysis can be applied in seismic data processing to describe the temporal properties of both noise and signal, and show the time-frequency structure of different kinds of waves in detail. The importance of this analysis is that it can reveal intuitively the frequency and energy distribution along time direction, thus providing a solid base for designing a reasonable scheme in data processing so as to obtain a high precision result. In this paper, time-frequency analysis method is presented, and suitable ways of selecting window functions are also introduced. Noise attenuation in frequency-divisional way demonstrates the value of time-frequency analysis and shows its effectiveness in high-fidelity processing.