Mutiscale Acoustic Impedance Inversion Using Poststack Seismic Data

In this paper, Multi-scale (multi-resolution) inversion is studied and applied. Multi-scale (multi-resolution) inversion is an inversion tactics that has been put forward in recent years. Its purpose is to increase inversion convergence rate and decrease the local minimum effect so as to search for the global minimum. Mutiscale inversion strategy in which original inversion problem is transformed into a series of cascade linearized problems has shown as a promising approach to nonlinear inversion problems. In this paper this idea is applied to acoustic impedance inversion using poststack seismic data. Firstly the necessity and significance for carrying this thesis was discussed, after summarizing various methods in impedance inversion area, the paper analyzed the advantages and disadvantages of each method. According to that, the model-based GLI algorithm which is much more mature is adopted to solve inversion problems in each scale. Then the paper researched the active theory in neoteric geophysics, wavelet transform, especially with its dividing-time and diving-frequence idea. In view of that, simultaneous multi-traces impedance inversion is achieved, to be more specific, in iteration process of each scale, we form the frequency-division wavelet residual traces resulting from real and synthetic seismograms, then with the use of K-L transform, the coherent contents of frequency-division section are taken to estimate wave impedance variations, thus preventing noise from joining in the estimation of impedance variations, so noise can be removed while impedance inversion is being performed. Theoretical model and real data inversion shows that the inversion results are reliable, and this method has the features of quick convergence rate, high inversion precision and strong ability of accommodation.