保持信噪比的相位分解反褶积方法研究

反褶积是目前提高地震数据分辨率的主要方法，传统的反褶积大都建立在地震子波最小相位的假设之上，这 与实际情况下地震子波更接近混合相位的事实不符。由于地层的吸收作用，实际地震信号中的高频成分较弱，很容易 受到高频噪声的污染。地震记录经过常规反褶积处理之后，在提高分辨率的同时，降低了地震记录的信噪比，信噪比 与分辨率的矛盾在实际地震勘探中十分突出。为此，提出一种基于相位分解并能保持信噪比的混合相位子波反褶积 方法，实现混合相位子波反褶积，在提高地震记录分辨率的同时，能自动控制高频噪声对地震资料信噪比的影响，使得 分辨率和信噪比达到相对和谐，提高地震记录表征地下构造的能力。

A Study on the Method of Keeping SNR in the Phase#br# Decomposition Deconvolution

The deconvolution technology is the major method to improve the resolution of seismic data. The traditional deconvolution is based on minimum phase wavelet assumption，but in fact，this is inconsistent with the actual situation of mixed phase seismic wavelet. Due to the formation′s absorption effect and the weak high frequency signal in the real seismic data， the high frequency signal is susceptible to high frequency noise pollution. The traditional deconvolution reduces the SNR of seismic record at the same time when improving the resolution，so the conflicts of the SNR and resolution of in practical seismic exploration work is serious. This paper puts forward a kind of mixed phase wavelet deconvolution method，which can keep the SNR of seismic data and at the same time finish phase decomposition. This method when improving seismic record resolution， can automatically control the high-frequency noise of seismic data to achieve harmony between the resolution and SNR，and improve the ability of seismic record characterization of subsurface structure.