Научная статья

Информатика, кибернетика и программирование

One cn distinguish three clsses of visuliztion systems. The first one consists of universl systems which include set of lgorithms for constructing wide rnge of typl representtions. For exmple wellknown systems PrView nd VS belong re of this kind.



29 KB

0 чел.


P.A.Vasev*, S.S.Kumkov*, E.Yu.Shmakov**

*IMM UrB RAS, Ekaterinburg       **UrFU, Ekaterinburg

Traditionally, scientific visualization is a methodology of converting abstract objects to some geometric images. This allows to observe visually results of numerical simulations. When some entity is visualized, the specific part of the process is choosing some certain two- or three-dimensional geometric representation of the abstract object and creating an algorithm for constructing this representation of the basis of data produced by the computational program. The instruments for manipulating the obtained geometric objects (moving, rotating, zooming, changing attributes – color, visibility, opacity – etc.) are usually the same for different visualizations.

One can distinguish three classes of visualization systems. The first one consists of universal systems, which include a set of algorithms for constructing wide range of typal representations. For example, well-known systems ParaView and AVS belong are of this kind. The second category is of systems for specific visual essences. Among them are IVS3D (for geo-information), VENUS (for molecular structures), VolVis (sparse three-dimensional arrays). Finally, the third class is formed by specialized visualization systems created for some certain project of, even, for some researcher.

Environments from two first classes are good because in many situations can allow to get quickly some result. But in practice, it often turns out that the algorithms built into these systems give objects, which satisfy user’s demands partially or even do not satisfy at all, and adding own modules for constructing geometric object is impossible.

The ideal variant for the end user is a system worked out with taking into account all his desires and needs. But developing such systems usually is started from the ground up, and to overcome all appearing problems it needs for qualified programmers, who usually spend a lot of time to recreate the user interface, which is repeated from system to system.

The authors have created a visualization system, which avoids the mentioned problems. It includes quite rich user interface and a powerful engine allowing to store a three-dimensional scene and work dynamically with it by means program interface (API). In the framework of the suggested system, a user to visualize a new type of objects should implement the corresponding construction algorithms as DLL-library of some prescribed format. The procedures put into the library read output files of corresponding computational program, construct respective geometric objects and add them to the scene.

The current version of the program is written in C# for Microsoft ,Net 4.0. It uses WPF window library of the .Net environment and the three-dimensional graphics library Media3D.

Thus, a creating a new visualization system stops to be a complex project including a lot of routine technical problems and becomes a tuning, adopting, and extending an existing modular system.