Seismic Signal De-noising Method Based On Independent Component Analysis

The seismic signals usually contain a variety of noise by the interference of externalconditions and formation reflection seismic exploration process. In order to eliminateinterference noise from seismic signals to get accurate information, the process of seismicsignal de-noising is very important. Seismic signals typically have non-Gaussiancharacteristics, and are independent with random noise statistically, in line with theapplication conditions of Independent Component Analysis (ICA), so this issue usesIndependent Component Analysis method for seismic signal de-noising. The main researchworks are as follows:Two main algorithms of ICA: FOBI and FastICA are introduced, and signal separationsimulation of the algorithm is presented. The result verifies ICA method’s validity, and gives aconfirmation for ambiguity of the amplitude in signal separation.Aimed at improving the shortcoming of easy to get the local optimal solution of InvasiveWeeds Optimization (IWO), this issue introduces niche ideas and put forward a Niche IWOalgorithm. The method improves the diversity of the population of IWO algorithm andenhances the ability of global optimization algorithm. The validity of Niche IWO algorithm isverified by the simulation of four standard function’s optimization and the ability of globaloptimization is enhanced.This paper combines the improved IWO algorithm with ICA criterion and gives animproved IWOICA algorithm. The effectiveness of IWOICA method is verified through theseparation of analog and hybrid mixed-signal. The result of IWOICA method seems betterthan FastICA in signal separation especially for hybrid mixed-signal separation.After givingthe IWOICA method, this paper uses the MATLAB simulation analog seismic signalsde-noising of dual-channel to verify the feasibility of the algorithm. And compare the resultwith the result of the F-X domain method, fourth-order blind identification (FOBI) and theFastICA algorithm to verify the superiority of the improved algorithm.In order to solve the limitations of single-channel in practice, this issue uses EmpiricalMode Decomposition (EMD) to reconstruct the second virtual signal as IWOICA two input source signals to realize the single-channel seismic signal de-noising and the result is inaccordance with expected。...