Hybrid Optimization Of Threedimensional Seismic Image Noise Reduction Structures Oriented Approach

Nowadays, due to the development of intelligent information technology, information system put forward higher requirements. Imaging system because of its intuitive, reliability has been widely applied in various fields. But because the limitations of the physical limitations and environmental conditions, the transmission medium system equipment is uncertainty, so that the image contains a lot of noise, the degradation phenomena that affect the application. It is very important to remove the image noise effectively.However, the data characteristics of different fields, make them requiring different image noise reduction technology. Among them, three-dimensional seismic images will have about a few features. First, a very large amount of data, usually a complete seismic work area data can reach tens Gb, some even hundreds of Gb. Second, the noise component is complex, due to the seismic data acquisition systems and the complexity of environmental data is collected, resulting Gaussian noise components containing complex components and non-Gaussian component. Third, there is a lot of need to protect the image of the edge of the texture, such as faults, pinch and so on. These structures are often indicative of oil and gas reservoirs important information, so the noise at the same time must be maintained.For these characteristics, the main strategy of this paper is to enhance the efficiency of the algorithm, to increase the ability of non-Gaussian noise denoising, enhancement structure retention characteristics. Emphasis on structure-oriented hybrid noise reduction method and optimizing noise reduction methods in the application of three-dimensional seismic image, carried out research. The main research work includes :1 、 There are some discontinuous faults, pinch out the three-dimensional seismic images and other information, these information are very important in geological studies. But the structure is not continuous-wave signal is weak, resulting in signal strength is relatively weak, so when denoising for the information we need to protect them. To solve this problem, this paper presents an improved structure-directing noise reduction method, the diffusion matrix by redesigning and continuity factor, to improve the convergence rate and structural algorithms to maintain performance.2 、In three-dimensional seismic image, due to system complexity of the image information and the natural environment, image noise components are very complex. There are not only the noise component Gaussian noise, as well as super-Gaussian noise and sub- Gaussian noise. The traditional approach is difficult to meet the various noise components effectively remove noise while maintaining a good image of some specific configuration information. To solve this problem, we propose a hybrid optimization based on three-dimensional seismic image noise reduction structures oriented approach. This method can be integrated structure-oriented structure to keep the noise reduction methods noise characteristics and hybrid optimization method can effectively filter out the noise characteristics of the various components.In this paper, the use of the proposed noise reduction methods and conventional noise reduction method for three-dimensional seismic images of the actual noise experiments. Comparing the two sets of results, the analysis found that the proposed method in the noise reduction effect than the conventional structure-oriented approach significantly improved noise reduction, and to maintain the structure also significantly strengthened. After this filtering, can achieve higher signal to noise ratio of seismic data, structural interpretation to improve the accuracy of the latter, the algorithm processing, artificial analysis, machine analysis and reliability is important.