The Research On Fast Volume Rendering Algorithm Based On GPGPU

As one of the major fields of scientific visualization, Volume Rendering is that generates two-dimensional image directly from three-dimensional data. This paper is based on typical algorithm of volume rendering, and is combined with the programmability features of GPU. The imaging process is speed up by GPGPU to finish the data-intensive computing of Volume Rendering. The main content and innovative points of this paper can be outlined as the following three aspects: Presenting a ray-tracing algorithm which is based on GPGPU; In frequency domain, introducing windowed-sinc function as a filter to imp lement high-order interpolation; Designing and developing a large-scale seismic data display and interpretation system.First, based on the study of volume rendering integral, the image-based volume rendering algorithm is analyzed. Since in the ray-tracing algorithm, the ray traversal data is more time-consuming,"SLAB"is introduced to calculate the intersection with the bounding box. After getting the entry point and the leaving point of each light, the color cumulative calculation is executed, which greatly improving the rendering speed. At the same time, in order to avoid software implementation of tri-linear interpolation, three-dimensional texture is introduced to implement resample by hardware. The experimental results show that this algorithm can obtain satisfactory interactivity without affecting the quality of the image.Second, in aspect of domain-based volume rendering algorithm, a new FVR algorithm is proposed through expanding Fourier Projection-Slice Theorem into High-Dimension and mapping the pipeline totally on GPU. Blackman-windowed function is used as reconstruction filter to implement higher-order interpolation and the reduction of samples is executed on GPU in parallel, which meets the architecture of heterogeneous multi-core. The rendering is accelerated by a factor of 7 after reducing the complexity of the algorithm from O(N3) to O(N2logN). Last, on the basis of the above, we design and develop a visualization system ofseismic data which is common and also easy to extend. The display and interpretation of different type of seismic data is implemented by plug-in. On this basis we solve some practical problems, such as large-scale data management, transfer function, volume clipping and bump mapping. By a variety of data validation, the rendering speed and the image quality of platform achieve very good results.