Mednarodni sejem inovacij SIIF 2012
Naslov inovacije: Differential Evolution for Parameterized Procedural Woody Plant Models Reconstruction within Environmental Framework to Visualize Emergent Artificial Forest Ecosystems
Inovator: dr. Aleš Zamuda
Področje inovacije: računalništvo in informatika
Opis (v angl.):
This invention presents an approach for reconstruction of procedural three-dimensional models of woody plants (trees). The used procedural tree model operates by recursively computing all building parts of a three-dimensional tree structure by applying a fixed procedure on a given large set of numerically coded input parameters. The parameterized procedural model can later be used for computer animation. Reconstruction of a parameterized procedural model from images is done by differential evolution algorithm which evolves this model by fitting a set of its rendered images to a set of given reference images. The comparison is done on pixel level of the images through the integration of distances to the nearest similar pixels. The obtained results show that the presented approach is viable for modeling of woody plants for computer animation by evolution of the numerically coded procedural model.  The reconstructed models are used within an environmental framework to visualize emergent artificial forest ecosystems. The environmental framework simulates woody plant forests which are used for visualization. A complex application software system develops and animates a spontaneous afforestation process within this environment. The system considers several environmental properties and combines computer animation with artificial life. The main goal of the presented software system is to use it in computer animation for synthesis of natural environments and visual analysis of their natural look credibility. The afforestation process is modeled as an ecosystem simulation, where trees struggle for survival based on several growth factors. A detailed description of the procedures for simulating tree growth and the factors that might influence tree growth is provided. All the tree growth simulation procedures and factors are biologically inspired. They have been defined mathematically in the paper by designing a bottom-up agent model which emerges the artificial tree distribution by mediating to the simulation. A flexible and adaptable procedural 3D model is used to visualize trees. Also, growth of individual trees is animated, from development of branch complexity to per-leaf precision, which allows a very realistic perception of the emerging ecosystem. The visualization of trees is sped up so that the models of trees have progressively lower-details proportional to the distance from a certain point of view. Locations and maturity of visualized trees are obtained from the ecosystem simulation results, and the afforestation process is animated over several centuries. The natural look of the artificial tree distribution is confirmed visually and statistically. Visually, it is confirmed from rendered sequences, and statistically, from graphs of tree species populations. Several patterns emerge permanently, such as the number of trees in the ecosystem simulation increasing exponentially and trees growing in communities. [2,3]
 A. Zamuda, J. Brest, B. Bošković, V. Žumer. Differential Evolution for Parameterized Procedural Woody Plant Models Reconstruction. Applied Soft Computing, vol. 11, no. 8, pp. 4904-4912, 2011.
 A. Zamuda, J. Brest. Environmental Framework to Visualize Emergent Artificial Forest Ecosystems. Information Sciences, 2013, vol. 220, pp. 522-540. DOI10.1016/j.ins.2012.07.031.
 A. Zamuda. EcoMod project code. Sourceforge http://sourceforge.net/projects/ecomod/.
 A. Zamuda. EcoMod project documentation. http://labraj.uni-mb.si/en/EcoMod
Odmevnost v medijih: