Seismic Volume Data Visualization And Analysis

Volume Rendering is one of the primary methods in scientific visualization field. The intrinsic information of volume dataset can be visualized interactively by this method. In medicine, meteorology and oil and gas seismic exploration application, volume rendering technology is widely used, and helps users to analyze the data and to find the rule or law concerning the data. With the emergence of various complex oil and gas discovering and forecasting problems, the traditional interpretation of seismic volume data visualization technology faces new challenges. Thus, it is vital important to study further the seismic volume data visualization in theory and practice.This thesis focuses on the key technologies of seismic volume data visualization and analysis, and proposed three novel seismic volume data visualization and analysis algorithms. All three new algorithms are integrated into our developed seismic volume data visualization and analysis system. Some seismic exploration data are tested by our system and the experimental results demonstrate the efficiency of our new algorithms.Firstly, we proposed an interval volume rendering algorithms by ray-opacity-modulation. Our algorithm can overcome complex specification of transfer function, feature occlusion, and view-dependent feature peeling which are the limitations of existing volume rendering algorithm. We adopted the interval volume to define the structures of volume data, in which the interval volume is defined by a sub-space with scalar values in a specific range and it can be used to describe visual feature with more accuracy. Transfer function can be automatically generated by the principle interval volume which is detected based on corresponding level of local data relevance. To assure the visibility of all interval volumes at once, we proposed a novel interval volume rendering equation based on ray interval profile analysis. By the application of our rendering equation, users can interactively achieve interval volume peeling rendering without view dependence.Secondly, to solve the problem of features missing by traditional high dynamic range seismic volume data quantization method, we proposed a feature preserved seismic volume data quantization method, and introduced a feature enhanced volume rendering algorithm. The seismic volume data quantization is an important preprocessing step before rendering high dynamic range seismic volumes on regular display screen. Base on the sub-range visual analysis of the high dynamic range seismic volume data, a series of features such as the fine structures, the local relevant structures and singularities structures of volume data are defined. Fine structures can be isolated by series iteratively edge preserved filter algorithms, local relevant structures are obtained by local relevance statistics, and then singularities structures are departed from them. We proposed guidelines and an optimization strategy of feature preserved quantization by analyzing the classification of features and their effects on volume rendering. To suppress the noise which may affect the quality of volume rendering of the scaled seismic volume data, we introduced a feature enhanced volume rendering algorithm.Thirdly, a method of robust multi-scale seismic horizon detection and visualization is introduced for more complex horizon detection and classification. Seismic horizons are the predominant geological features which indicate changes in rock properties, so horizon detection is more important to seismic interpretation. But many existing horizon detection algorithm cannot work well due to the high noise of seismic data, and discontinuity points often results in ambiguous horizons detected. To suppress the noise and enhance the continuity of horizon, we proposed a multi-attribute based filtering algorithm to enhance the structures of seismic volume data. Our multi-attribute based filtering algorithm defined the filter weight by the intensity, space distance and local structures of seismic volume data. The filter window of our multi-attribute based filtering algorithm can be rotated along dip to enhance the continuity of horizon, translated with coherence to keep the discontinuity features. The filter weight can be adaptively adjusted with chaos due to the noise level of region to achieve different filter effects. We proposed a flatness-based multi-scale horizon detection method for the enhanced seismic volume data by our multi-attribute based filtering algorithm. The method can adaptively adjust the scale of detection by local flatness and achieve horizon classification. To observe the detected horizons, we implemented the multi-volume visualization technique and proposed a method to visualize the processing of horizon detection. Finally, we developed a seismic volume data visualization and analysis system that integrates the above feature enhanced visualization and horizon detection methods. The system also provides various functions, such as seismic attributes estimation, multi-attribute volume visualization and seismic slice processing, etc.The research of this thesis explored seismic attributes calculation algorithms, volume rendering methods and graphics hardware techniques, and provides the new methods for complex seismic interpretation. A number of experimental results demonstrate the efficiency of our proposed methods.