Font Size: a A A

Three-dimensional Data Field Body To Draw Several Problems

Posted on:2002-08-28Degree:DoctorType:Dissertation
Country:ChinaCandidate:W W SunFull Text:PDF
GTID:1118360032451215Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Visualization for three-dimensional data fields is commonly used in many applications wherevolume rendering is widely applied. However, volume rendering is inherently time-consuming.In this paper, by investigating the process in which researchers study data field, we proposesome new methods to improve the efficiency of studying data field through volume rendering.The research works conducted in this paper are sunnnarized as follows:l) In object-order volume rendering for irregular data fields, the cells occluded can beculled directly so as to reduce the amount of data and accelerae the rendering. In this paper, weuse an average opacity buffer by convolving the opacity buffer with a box filter of the size whichis caculated in preprocessing, each pixel in this average opacity buffer mpsents the averageopacity of all pixels within a neighborhood of the dimensions of the projection bounding boxwith the size of box filteL Thus, the average opacity buffer enables us to detenninate if a cell isfully occluded, accordingly culled, or ifpanially occluded, accordingly proected.2) Stereoscopic images can help researchers comprehend complex data fields by employingdepth information. But it would take a long time to render stereoscopic images. In this papef, byprojecting onto right image plane the intermediate images composited from the images onsuccessive slicing planes, while rendering the left image based on a series of parallel slicingplanes[Yage96], the computation of composition in rendering the right image can be reduced,and the efficiency of stereoscopic rendering can be improved.3) It is an effective way to st'Udy interesting contents by combining and rendering variouscontents so as to comprehend them and their relations with each other In this papef, a new datastructure is proposed to allow an efficient combinative volume rendering convenientlyperfOrmed. By creating indexed data structures for each content, only interesting contents arerendered so as to prevent the comPutation for the uninteresting pats. In addition, the cells ofeach content are handled in order to make it easy to queue uP the cells of the interestingcontents.4) Volume navigation can overcome the drawbacks of excessive comPutation for full datarendering. In this paper, projection method is taken to render hages based on surface extractedon the fly, so the rendering is accelerated. Meanwhile, local OPerahons are suPPOrted on data tooptimize visualization of local featUres.
Keywords/Search Tags:visualization, volume, rendering, object ordef, image ordef, culling, occlusion, stereoscopic rendering, combination, data smicture, navigation.
PDF Full Text Request
Related items