| Reverse time migration(RTM)is currently one of the most effective techniques for imaging complex structures.However,RTM based on two-way wave equation suffers from heavy computations and large storage requirements.Meanwhile,time dispersion in wave equation simulation will cause depth and phase deviation in imaging.Therefore,enhancing computational efficiency,reducing storage burdens,and decreasing the effects of time dispersion will be of vital significance to the development of RTM.To deal with the large storage and highly frequent I/O in RTM,this study combined random boundary condition and absorbing boundary condition(termed hybrid random condition as below)to perform wavefield reconstruction.The random boundary was prescribed by controlling the size and shape of the random particles,and the reference value of random particles setting was given according to the scattering theory,which realized the reduction of storage and the improvement of computations efficiency.The method was also extended to the wavefield reconstruction in elastic reverse time migration.Based on the first-order velocitystress equation,the time dispersion caused by the second-order time difference of staggered grids for both acoustic and elastic wavefield extrapolations were analyzed.The method of time dispersion correction was applied to elastic reverse time migration to obtain images with corrected imaging depth and phase even in the case of large time steps.This study utilized simple and standard models to test the proposed methods.Numerical simulation shows that wavefield reconstruction using hybrid random boundaries can efficiently reduce the amount of storage without affecting the imaging results.Due to less I/O,the calculation efficiency is also significantly improved.Time-dispersion corrected RTM using large time-steps can effectively reduce the errors in imaging phase and depth,indicating relatively enhanced computational efficiency. |
