Abstract: Recent years, to improve seismic prospecting records and increasing signal-to-noise ratio, some filtering techniques have been presented, such as the chaotic vibrator method (CVM) based on chaotic theory, the hyperbolic time-distance relation filtering (HTDF) based on the least-squares algorithm, the time-frequency peak filtering (TFPF) based on the time-frequency analysis and Wigner-Ville distribution, the technique decreasing multiples based on independent component analysis (ICA). The CVM is applied to ascertain the location of events with strong random noise. The HTDF is the processing technique following CVM to obtain a clear event recovery from the seismic prospecting records corrupted by random noise. Tests of these methods on both simulation experiments and practical seismic data processing show that the TFPF technique can successfully decrease the random noise. The technique to attenuation multiples based on ICA can differentiate the output signals. However, above four techniques have still some questions which require to be solved, such as the computation efficiency, the assistant processing, the supplement-frequency method of amplitude spectra of the filtering factors, the adapting time-window technique, the systematism of processing performance and so on. Above processing techniques are a part of theory and method in seismic data processing.

Key words: chaotic vibrator filtering, hyperbolic time-distance relation filtering, the time-frequency peak filtering, decreasing multiples