Joint Seismic Data Reconstruction Based On The Accelerated Bregman Method And The Threshold Iteration Method

Due to the limitation of terrain conditions or economic costs,seismic data is usually irregularly sampled along the spatial direction,resulting in a serious loss of data.Therefore,the reconstruction of missing channel seismic data is an important link in data processing,but most reconstruction algorithms have slow convergence speed and high computing cost,which is difficult to meet the requirements of massive data processing.To this end,this paper uses multiscale multidirectional curved wave transformation as a sparse basis,introduces the accelerated linear Bregman method,combines it with the threshold iterative method,and proposes a fast joint reconstruction method.Because accelerated linear Bregman method can get more effective signals from unthresholding curved wave coefficient,so in early iteration convergence fast,late unthresholding curved wave coefficient into more noise,resulting in reconstruction accuracy is poor,and threshold iteration method requires threshold curved wave coefficient,early iteration filtering most of the effective coefficient,so convergence is slow,but later can recover weak effective signal,so the reconstruction accuracy is high.In order to effectively use the advantages of the two algorithms,this paper proposes the linear and exponential range of linear Bregman method in the early iteration,late iteration threshold iteration method,thus making the joint algorithm iteration speed.Based on this,a new acceleration factor is proposed to further accelerate the iteration efficiency,and the iteration accuracy is greatly improved to meet the requirements of massive seismic data reconstruction.Meanwhile,in the joint process,different threshold factors and threshold formulas are discussed,and finally the soft threshold function is adopted,and a new exponential threshold parameter formula is introduced.Theoretical simulation and practical data processing show that: compared with the threshold iteration method,the linear Bregman method,the accelerated linear Bregman method and the traditional joint method,the accelerated joint method proposed in this paper is fast and the reconstruction effect is obvious.