Research On Seismic Data Visualization And Intelligent Fault Recognition Method

Seismic data visualization and fault recognition are important sections in seismic interpretation,and have been extensively studied in oil and gas exploration area,but have not been well solved.The development of computer technology has led to the development of this research to a new stage.The purpose of seismic data visualization and fault recognition is to reveal the geological structure information such as horizons and faults in seismic data to facilitate seismic interpretation.The existing seismic data visualization and fault recognition methods provide an important reference for the seismic interpretation work to a certain extent,but there are still some shortcomings.There are two main disadvantages of the existing seismic data visualization method:(1)The linear coloring method results in a lower contrast of the visualization results,and the geological structure is difficult to distinguish;(2)The ability to reveal the micro seismic horizons and the micro fault structures is too weak.There are three main disadvantages of the existing fault recognition method:(1)the accuracy of the fault is low and it is easy to mislead the seismic interpretation work;(2)the time consumption is too large;(3)the recognition ability for the multiple faults,especially multiple cross faults is too poor.From the perspective of computer applications,this paper focuses on the three-dimensional visualization of the two-dimensional seismic section data and the fault recognition methods.The purpose of our work is to solve the shortcomings of the existing methods,so that the ability of seismic data visualization to reveal the micro geological structures,the accuracy and effectiveness of the fault identification are improved.The research contents include two aspects: seismic data visualization and fault intelligent recognition.The main research work and innovation of this thesis are as follows:1.A three-dimensional visualization method for two-dimensional seismic section is proposed to improve the ability to reveal the micro geological structure.The method mainly consists of two aspects.For the vertices coloring in the seismic mesh,the color component of palette was filled in piecewise exponential progressive style at first,then use the texture coordinates which have been mapped from the seismic magnitude to sample the palette for the coloring of each mesh vertex in pixel shader of the GPU;for the tessellation of a sample,the dip style of each sample is calculated in the pixel shader and is rendered to a dip texture,and the vertex index is filled according to the dip style so that the seismic sample can be tessellated adaptively;Finally,the experiments validated the effectiveness of the proposed method.2.A fault recognition method based on ant colony optimization is proposed to improve the accuracy.This method used the seismic section as a three-dimensional terrain,first,horizons are extracted by the connected component labeling algorithm;then,the horizontal endpoints of each horizon are used as the discontinuties;after that,the seismic section is divided into several rectangle blocks and the top endpoint and the bottom one in each block are regarded as the nest and food respectively for the ant colony optimization algorithm;when the ants starting from the nest to find food,they crawling on the three-dimensional terrain generated by the seismic section,and the projection of the shortest three-dimensional path between the nest and the food found by the ants in each rectangular block is the fault.The experiments on the actual seismic data verified the effectiveness of the method.3.A fault recognition method based on the affinity propagation clustering algorithm is proposed to improve the correct rate of fault.This method is proposed based on the affinity propagation clustering algorithm,the horizon discontinuities are located by the connected component labeling method first;then the similarity is defined by the horizontal gradient difference and Euclidean distance between discontinuities;after that,the discontinuties are clusterted through the iterative transmission of the responsibility and the availability messages between the clustering center and the points to be clustered;finally,the PCA method and the least squares curve fitting method have been used to generate the fault based on the discontinuties of the same class successively.The experimental results show that the accuracy of the method is high,about 90%.4.A fault recognition method based on adaptive clustering Hough transform is proposed to improve fault precision.This method is proposed to solve the problem that multiple faults can't be recognized correctly by the existing fault recognition method based on Hough transform.First,edge detection was performed on the seismic coherence image and Hough transform of the edge image was performed to detect line segments;Then,the line segments were clustered adaptively according to their inclination angle and position to obtain a more complete line segment;Finally,the position of each point on the complete line segment was optimized according to the original seismic image to generate a more accurate fault.The experimental results on the real seismic data indicate that multiple faults were recognized correctly by the proposed method,and the peak signal noise ratio of the result was improved by 10%.5.A fault recognition method based on adaptive cluster-wise linear regression is proposed to improve the accuracy of fault.This method recognize the faults based on the adaptive cluster-wise linear regression algorithm.First,the endpoints of the line segments detected by Hough trans-form are used as key points;then the key points were clusterted by the adaptive cluster-wise linear regression algorithm based on the linear relationship between them;finally,a fault is generated from each class of key points.The experimental results on the real seismic data confirmed the effectiveness of the method.In this paper,we focus on the enhancement and extraction of the information in seismic data,and studied the three-dimensional visualization of the two-dimensional seismic data and the intelligent fault recognition method.The methods proposed in this paper can help the geological researchers to analyze the seismic data more effectively,which have a good application value and practical significance for oil and gas exploration,oil extraction and agricultural production.