Seismic Data Denoising By Using Anisotropic Diffusioin Filter

Our purpose is denoising the noise of two-dimensional seismic data, and improve the quality of the two-dimensional seismic data. We got an better resultes by using the method based on nonlinear anisotropic filtering. This filter is based on partial differential equations, and allows us to use the details and fault information. By calculating the structure tensor, we calculated the direction of the local geological layer information, and use of the structure tensor information to build the diffusion tensor, and the diffusion along earthquake. We called this method: ANDF. In the simulation and experiment we use seismic data, and have got effective results.In the first chapter, we mainly talked about the background of the subject, as well as the status and stage of research studies. And analysis the pros and cons of the main research methods, provide a realistic demand for our topic is expanded. In the second chapter we discuss about the seismic data security and major development side filtering defects. And discusses partial differential method for image processing. In the final chapter, we describe the criteria that used in this article and the calculation formula.The third chapter we introduce the anisotropic diffusion filter, and elaborated the diffusion physics background and scale space. And we elaborated that we discussed filters in this article are based on partial differential equation. After the introduction of the anisotropic diffusion filtering technology,we introduce that the different diffusion tensor,the different eigenvalues, and we got the results are different. And finally we put forward own diffusion tensor.Main work done, is a senior scholar diffusion tensor were used to achieve, understand their strengths and weaknesses. And the existing basis, to improve, to adapt to the seismic design of diffusion tensor image processing design. Shortly after, on a five-point difference diffusion equation discrete, get the iteration scheme. In the analog image is found, the seismic image noise suppression diffused better, and better seismic images tomographic reserved.