Research On Method Of Volume Rendering About Mass Seismic Data

Volume rendering on 3d dataset has been widely used in the fields of medical, fluid physics, meteorological, geological exploration etc, and the visualization of mass 3d dataset has drawn attention of the field of academia and industry. Usually, there are three idea of the realization of three-dimensional dataset visualization:the first is the thought of parallel, of which multi-machine of multi-GPU is used to handle the problem; Secondly, the thought of compression, of which a compression method is used to reduce the amount of data; the last one is the idea of multiresolution, of which a LOD visual technology solutions and processes is used. The compression volume rendering based on tensor approximation which uses the thought of high order PCA, implements the efficient compression of 3d dataset. For seismic data compression processing, the fact that the factor matrixes are exclusive (not shared) leads to the problem of low compression effect and low reconstruction efficiency. In this paper, aiming at the problem of the above-mentioned two aspects of research, the main results were as follows:(1) A global factor matrix based tensor approximation method was proposed. In the traditional tensor approximation method, each block has its own factor matrix, which lead to data redundancy and a low real-time reconstruction efficiency. In order to solve this problem, this paper presents a tensor approximation method based on global factor matrix. A global factor matrix for the all bricks was used to reconstruction, considering each brick of data is obviously has a certain similarity, Simulation shows that this method can be consistent with the traditional method of compression ratio and peak signal-to-noise ratio, but has higher efficiency of refactoring;(2) A factor matrix clustering based tensor approximation method was proposed. In a method, we select a set of common factor matrix to enhance the compression effect, caused the reconstruction effect of loss, under the same compression rate peak signal-to-noise ratio and the traditional methods are basically identical. Considering the similarity between different block size differences, is proposed in this paper, based on the tensor factor matrix clustering approximation method, on the basis of the previous method, factor matrix of each partitioned clustering, to belong to the same cluster factor matrix block, using the same factor matrix refactoring. Simulation results show that compared with the previous methods, this method can guarantee the basic under the same compression ratio, significantly increase peak signal-to-noise ratio, while still maintaining high efficiency of refactoring.The place on put together is narrated, in this paper, in the compression of seismic data volume rendering tensor decomposition and reconstruction of independent factor matrix caused by the problem of low efficiency of compression ratio and reconstruction, puts forward and implements the effective solution algorithm, and implements the method in this paper for the simulation of the tensor approximate compressed volume rendering software platform. Compared with the original method, the method can get higher under the same compression ratio of peak signal to noise ratio, has the very high practical value.