Study On The Method To Suppress The Random Noise Of The Geological Boundary

In the process of collecting in the field,the underground geological conditions are often complex and changeable,and the external random disturbance will affect the resolution of seismic data.Because of the irregularity and complexity of the underground geological body and often distributed in the complex structure and other reasons,there is a large difference in the effective energy of reflection information,which makes it difficult to get the balance of the seismic data denoising and fidelity requirements.In this paper,through the analysis of different geological body edge structure information of seismic image data to construct random noise in multiscale diffusion tensor analysis,and changeing the diffusion rate in the flat region and the region with abundant information,so as to achieve the balance between seismic data denoising and fidelity requirements according to the idea of image edge enhancement.The main contents of this paper include the following aspects:(1)the analysis of the causes and types of seismic random noise in the field acquisition and its influence on seismi;(2)The gradient analysis of seismic data structure is used to solve the seismic data structure information.(3)All kinds of geological models are established to analyze the characteristics of reflection and to identify the noise of geological bodies.(4)In this paper,the basic theory of partial differential equations and the method of multiscale analysis are introduced;The nonlinear anisotropic diffusion filtering of linear isotropic diffusion equation and P-M model are introduced;The denoising ability of various filtering methods in theoretical model and actual seismic data is analyzed.Compared with the traditional filtering method,it can be found that the method based on anisotropic diffusion has better denoising effect in the seismic data discontinuity information.Processed seismic data provides a good foundation to shorten the data of seismic data interpretation and analysis of the cycle for the late structural interpretation,breakpoint identification,horizon tracing,extended high resolution processing,and geometric attribute extraction.