Computation of Hessian Matrix Based on Dynamic Programming Traveltime Field and Its Application
The seismic data inversion imaging is an extension of classic migration imaging within the least squares inversion framework.Its core lies in eliminating the impact of the Hessian matrix in the migration imaging system to achieve higher-resolution imaging results.The immense scale of the Hessian matrix makes its efficient computation crucial in advancing seismic data inversion imaging methods.This paper compares the migration of a reflection wave with inversion imaging,deriving the expression and simplified form of the Hessian ma-trix.By analyzing the expression,it identifies the traveltime field as a key issue in Hessian matrix computation.Next,it proposes an approximation of the Hessian matrix using the point spread function(PSF)and introduces dynamic programming for traveltime computa-tion,obtaining the PSF representing the traveltime field.This approach achieves efficient representation of the first-order Hessian matrix and analyzes the properties of the computed PSF while comparing computational efficiency.Finally,utilizing the PSF computed based on traditional migration imaging results,it corrects the resolution in the imaging domain and tests and applies the effectiveness of the computed Hessian matrix.Numerical experiments demonstrate that the Hessian matrix computes efficiently and stably based on dynamic pro-gramming for the traveltime field.By comparing imaging results before and after Hessian matrix resolution correction,it validates the effectiveness and application potential of the de-veloped efficient Hessian matrix computation method.
NETL
NSTL
万方数据
维普
