Automatic L-System Discovery for Virtual Plants by Branching Pattern Analysis of Unfoliaged Trees

作者：Qu Hongchun^*; Zhu Qingsheng; Fu Hegang; Zeng Lingqiu; Guo Mingwei; Lu Zhonghua

来源：Journal of Computational and Theoretical Nanoscience, 2010, 7(5): 900-910.

DOI：10.1166/jctn.2010.1437

• L-System
• Virtual Plants
• Automata
• Plant Branching Pattern Analysis
• Unfoliaged Trees

Over the last decades, L-System has been widely used as a powerful tool in virtual plants modeling, particularly, in plant branch constructing. However, it is really difficult to manually develop a L-System for a given plant species only depending on imagination or experience. In this research, a novel approach of automatic L-System discovery by means of branching pattern analysis of unfoliaged trees is presented to address this issue. In this proposed approach, three steps are involved for L-System extraction. Firstly, image processing as well as pattern recognition methods are employed to recover topological and geometrical information for growth units and metamers from multiple images of unfoliaged trees. Secondly, Markovian methods are used to further analyze data which have been extracted in the first step for capturing the internal relationship among plant elements (on both scale of growth unit and metamer). Thirdly, the L-System has been generated throuth the running of a Bidimensional Hierarchical Automaton (BHA), which is constructed from the analysis result of the second step for describing plant branching structure. Experimental results show that our approach can discover L-System for unfoliaged trees effectively.

• 出版日期2010-5
• 单位重庆大学