Research Of Engineering Seismics Data Denoise Disposal Technology

Because the engineering seismic exploring layer is shallow , working condition and background are complicated ,the results are badly disturbed. How to remove kinds of noises in engineering seismic data processing and how to improve the resolution of seismic data are main task in geophysical exploration .Random noise has no uniform regularity ,it is showed in whole record paper and very hard to remove. Therefore how to effectively get rid of random noises is a goal which many people want to reach. Wavelet Transform is a method of time-scale (time-frequency), which has multi-resolution characteristics. It can express partial characters in time and frequency domain.This article tries to use the new method of wavelet transform to remove disturbance in engineering seismic data processing.First of all ,the article gives a brief introduction about all kinds of methods which are used to remove coherent noise and random noise ,and analyses their advantages and disadvantages according to the features and formation reasons of seismic signal noises .Then it describes the theory about wavelet transform in detail and the principle of discrete dyadic wavelet transform and the construction of orthogonal wavelet are discussed on the base of multi-dimension analysis tools. At the same time this article gives the algorithm of decomposition and reconstruction based on multi-resolution .Because of the frequency division in wavelet decomposition , we can use wavelet transformation to remove noise. The principle and method of wavelet threshold to remove noise are given and it is very important to make sure of the threshold in removing noise with wavelet transformation. When model signal use multi-layer decomposition to remove noise ,the results are different with different threshold .And the best threshold needs to obtain according to the calculation and setting in different kinds of threshold.According to the characters of wavelet transformation , on the plat of Matlab wavelet function toolbox,this article processes single track and whole profile data about engineering seismic exploring reflection and shallow seismic image ,uses the methods of de-noising by threshold filter in wavelet domain and de-noising method by threshold filter in wavelet packet domain to handle random noises in theory models and practical data,and analysis treatment effect. The experiment shows that performance of the method based on wavelet transform to remove disturbance in engineering seismic data is good and effectiveness. After the imaging system brought forward by this thesis being optimized, it can be directly used in the practical work.