The Earthquake Data Stack Went Voice Methods Research

Denoising is very important in seismic data processing. It will not only influence the effectiveness of other processes, but also have direct effects on the interpretation and the application of seismic data. A lot of work has been done in this field and many good experiences have been obtained, especially in after-stack data processing area. As the condition from surface to underground becomes more and more complicated with the development of seismic exploration and the data processing changes from after-stack to pre-stack, the effectiveness of the existing denoising methods are not satisfactory and more effective pre-stack denoising methods are explored.Complex geology condition brings about the complication of noise. Very strong relative interference, heavy surface wave, and even the interference noise caused by them are often found in seismic data of some complex area. In some cases, the reflection waves are mixed with them and can hardly be distinguished from them. In addition, first arrival is often interfered with various noises, which makes first arrival inaccurate to pick or even impossible to pick. Although a lot of researches have been done on the denoising of relative interference and surface wave, only a few of them have been used successfully in industry. There is even less researches on the denoising and making out the break time of pre-processing of first arrival. Focusing on these problems, this paper analyzed the existing theories, explored their industrial applications and made some progress in this area.Based on the study of automatic-trace theory, SVD theory and a lot of model tests, this paper presents a new denoising method and its application technology to eliminate relative interference. This new approach combines automatic-trace theory and SVD method together. It integrates the adjustment of covariance to avoid the lost of valid wave. It also makes denoising work in places having relative interference and not work when such kind of noise is not presented.For the pre-processing of first arrival, this paper presents a method that...