Streaming HLODs: An out-of-core viewer for network visualization of huge polygon models

S. Gumhold und S. Noll (Editoren)
In: Computers and Graphics (Feb. 2004), 28:1(43-50)
 

Abstract

Trimmed NURBS models are the standard representation used in CAD/CAM systems and accurate visualization of large trimmed NURBS models at interactive frame rates is of great interest for industry. To visualize the quality of a surface several techniques like isophotes, reflection lines, etc. are used. Most existing approaches transform the NURBS surfaces into a fine polygonal representation and build static levels of detail from this representation. This polygonal approximation together with its normals are adjusted in a semi-automatic procedure to achieve the desired visual fidelity during visualization. Since this approach allows only for a fixed maximum accuracy and does not support deformable models, another more recent approach is to keep the NURBS representation and generate view-dependent LODs on the fly up to the currently required preciseness.

However, so far this approach took only into account the geometric error of an approximation and thus neglected the various illumination artifacts introduced by the chosen (possibly view-dependent) triangulation. Although this problem can be solved by using normal maps, the resolution of the normal maps again limits the accuracy. Furthermore, the normal map generation requires a preprocessing step which prevents the support of deformable NURBS models. In this paper we present a novel automatic tessellation algorithm that considers the illumination artifacts and is well suited both for the generation of static and dynamic LOD schemes with guaranteed visual fidelity. Our new method is also capable of high quality visualization of further attributes like curvature, temperature, etc. on surfaces with little or no modification.

Stichwörter: compression (coding), distributed/network graphics, viewing algorithms

Bibtex

@ARTICLE{guthe-2004-streaming,
     author = {Guthe, Michael and Klein, Reinhard},
     editor = {Gumhold, S. and Noll, S.},
      pages = {43--50},
      title = {Streaming HLODs: An out-of-core viewer for network visualization of huge polygon models},
    journal = {Computers and Graphics},
     volume = {28},
     number = {1},
       year = {2004},
      month = feb,
  publisher = {Elsevier},
   keywords = {compression (coding), distributed/network graphics, viewing algorithms},
   abstract = {Trimmed NURBS models are the standard representation used in CAD/CAM systems and accurate
               visualization of large trimmed NURBS models at interactive frame rates is of great interest for
               industry. To visualize the quality of a surface several techniques like isophotes, reflection lines,
               etc. are used. Most existing approaches transform the NURBS surfaces into a fine polygonal
               representation and build static levels of detail from this representation. This polygonal
               approximation together with its normals are adjusted in a semi-automatic procedure to achieve the
               desired visual fidelity during visualization. Since this approach allows only for a fixed maximum
               accuracy and does not support deformable models, another more recent approach is to keep the NURBS
               representation and generate view-dependent LODs on the fly up to the currently required preciseness.
               
               However, so far this approach took only into account the geometric error of an approximation and
               thus neglected the various illumination artifacts introduced by the chosen (possibly view-dependent)
               triangulation. Although this problem can be solved by using normal maps, the resolution of the
               normal maps again limits the accuracy. Furthermore, the normal map generation requires a
               preprocessing step which prevents the support of deformable NURBS models. In this paper we present a
               novel automatic tessellation algorithm that considers the illumination artifacts and is well suited
               both for the generation of static and dynamic LOD schemes with guaranteed visual fidelity. Our new
               method is also capable of high quality visualization of further attributes like curvature,
               temperature, etc. on
               surfaces with little or no modification.},
       issn = {0097-8493}
}