A multi-channel deconvolution method for self-adaptive signal recognition

Deconvolution plays a critical role in enhancing the resolution of seismic data.However, conventional deconvolution methods, though boosting the high-frequency components of seismic signals, amplify the energy of high-frequency noise, thereby reducing the signal-to-noise ratios(SNRs) of seismic records after deconvolution.The contradiction between resolution and SNRs restricts the ability of existing deconvolution methods to characterize thin-layer structures.Hence, this study proposed a multi-channel deconvolution method for self-adaptive signal recognition.The method extracted seismic signal recognition operators from raw seismic data.It introduced them as spatial regularization constraints into the objective function of multi-channel deconvolution, somewhat achieving high-resolution processing with self-adaptive signal recognition capabilities.Based on the spatial predictability of seismic signals, their recognition operators were estimated and extracted directly from seismic data, demonstrating high adaptability to seismic records.As indicated by the test analysis of the model and actual data, the proposed method can effectively suppress the amplification effect of high-frequency noise during deconvolution, thus improving resolution and maintaining the SNRs of seismic records.