Acta Univ. Agric. Silvic. Mendelianae Brun. 2009, 57, 79-88
Published online 2014-10-14

3D visualization and finite element mesh formation from wood anatomy samples, Part II – Algorithm approach

Petr Koňas

Ústav nauky o dřevě, Mendelova zemědělská a lesnická univerzita v Brně, 613 00 Brno, Česká republika

Paper presents new original application WOOD3D in form of program code assembling. The work extends the previous article “Part I – Theoretical approach” in detail description of implemented C++ classes of utilized projects Visualization Toolkit (VTK), Insight Toolkit (ITK) and MIMX. Code is written in CMake style and it is available as multiplatform application. Currently GNU Linux (32/64b) and MS Windows (32/64b) platforms were released. Article discusses various filter classes for image filtering. Mainly Otsu and Binary threshold filters are classified for anatomy wood samples thresholding. Registration of images series is emphasized for difference of colour spaces compensation is included. Resulted work flow of image analysis is new methodological approach for images processing through the composition, visualization, filtering, registration and finite element mesh formation. Application generates script in ANSYS parametric design language (APDL) which is fully compatible with ANSYS finite element solver and designer environment. The script includes the whole definition of unstructured finite element mesh formed by individual elements and nodes. Due to simple notation, the same script can be used for generation of geometrical entities in element positions. Such formed volumetric entities are prepared for further geometry approximation (e.g. by boolean or more advanced methods). Hexahedral and tetrahedral types of mesh elements are formed on user request with specified mesh options. Hexahedral meshes are formed both with uniform element size and with anisotropic character. Modified octree method for hexahedral mesh with anisotropic character was declared in application. Multicore CPUs in the application are supported for fast image analysis realization. Visualization of image series and consequent 3D image are realized in VTK format sufficiently known and public format, visualized in GPL application Paraview. Future work based on mesh improvement through mesh error statistic, image acquisition and thresholding improvement by more sophisticated filters together with code optimization for fast image analysis is discussed. Also fractal characteristics classification on microscopic scale level is taken into account for further work.


10 live references