Based On A Large-scale Three-dimensional Seismic Data, Interactive Rendering Technology

Visualization in Scientific Computing is a kind of technology, by which the data generated in computing science and engineering can be changed into images or graphics and displayed on the computer screen. It's based on principles and methods of computer graphics. Seismic exploration is the primary means of oil and gas exploration, applying visualization technique to seismic data can greatly improves the efficiency of oil exploration, along with major economic and social benefits. Generally visualization can be classified into surface rendering and direct volume rendering. Compared with surface rendering, volume rendering has many advantages such as no division; visualize internal features of data field. However, as the calculation of volume rendering must process all of the data, for large-scale 3D seismic data direct volume rendering is a double-intensive problem of computing and the data. So how to improve the imaging speed to achieve interactive volume rendering is a focus nowadays. In this paper, considering the characteristics of 3D seismic data, we improved the traditional shear-warp algorithm, and realized the interactive volume rendering of 3D seismic data.Firstly, we discuss the classic volume rendering techniques, including advantages as well as disadvantages, in which the Shear-Warp algorithm is a point, and propose a shear-warp algorithm for 3D seismic data based on cube data structure considering the special characteristics of the 3D seismic data; Comparing with the acceleration of Octree and Relative Length Encoding (RLE), the algorithm constructed combination between voxels by cube structure; then classified them by the characteristics of seismic data. In the drawing process, we can use the binary tree index to quickly find the classification results, by avoiding the interpolation of the equivalent volume element calculations through the results and skipping transparent voxels, we improve the rendering speed greatly. While for the ripple distortion of traditional Shear-Warp, we designed a cube element-based resampling method, by which the sampling points increased and the accuracy of the image also improved.Secondly, In terms of rendering speed, we take advantage of multi-threading technology to accelerate the realization of the algorithm on the advantage of multi-core processors in parallel computing. In the final stage of generating image we design a look-up table to reduce the number of fusion operations and improve the rendering speed once more.Finally it comes to the completion of the 3D seismic data visualization system by using of QT and OpenGL. In the system module of volume rendering, we realize the interactive volume rendering of seismic data through the way of updating by using background thread. In the system module of brick file, In order to solve the problem of data reading speed through SEG-Y file, we propose a brick file format and through this kind of file we realized the interactive surface rendering of 3D seismic data in any direction in the system module of surface rendering.