Study On Method Of Improving Seismic Imaging Resolution

Deconvolution is one of the three basic technologies in seismic data processing, it is very important in all of the data processing flow. The result of deconvolution is relevant to the method applied to deconvolution and the hypothesis given before. Generally we suppose the wavelet is mini-phase, while in practical seismic exploration, the wavelet generated by dynamite is mixed-phase after earth filtering, thus when we obtain inverse filter and convolve it with the real data, it may lead to great differences even false result. Therefore many geophysicists are dedicated in studying new deconvolution method applicable to mixed-phase record.Statistical spiking deconvolution algorithm is devised in this paper, to breakthrough the restriction of mini-phase record in spiking deconvolution method, a new method called wavelet deconvolution is induced, this method adapts to mixed-phase record. The algorithm of this method is derived in detail, and the chart of program is given. The key challenge to obtain satisfactory result when applying this method to field record is how to extract precise wavelet. Statistical Hilbert transformation is used to obtain wavelet in this paper, theoretically, this method is only available to extract wavelet form mini-phase record, in order to make this method available to mixed-phase record, exponential decay is applied to data of computation window to transform it to the mini-phase, when the wavelet is obtained, make it divide the quantity that have been decayed before; meanwhile, multi-trace statistical method could remove the disturbance of surface wave and high noise, after that more precise amplitude spectrum could be received, so the wavelet which is in accordance with field record could be obtained.The wavelet deconvolution is superior to the spiking deconvolution on the synthetic data and real data examples shown in this paper, the former is available to any kind of phase data and the result is better.