Geometry Reconstruction of Plants

Abstract

Synthetic modelling and reconstruction of the geometry of real plants have been a center of attention of research for decades. Due to the complex architecture and growth pattern of plants, accurate modelling of the plant geometry is an extremely challenging task. Although realistic modelling of plants are widely studied in the context of computer graphics research, it also has profound impact on the biological study of plants. In order to perform various types of simulation studies under different environmental conditions and in understanding the physiology of plants in more details, synthetic models can be an extremely useful tool. Synthetic modelling approaches can be broadly categorized into three types. The first type is the rule based procedural modeling approach, which does not account the real data into consideration. The second type of approach (also known as data driven modelling) performs modelling based on the real data obtained from 3D acquisition procedures. The third type of approach is interactive, which is based on user assistance. In this chapter, we focus on the modelling of the second category and revisit the recent state-of-the-art techniques performing reconstruction of plant geometry from real data. The algorithms can be classified into different interlinked categories, which constitute the general pipeline of geometry reconstruction in data driven modelling framework. In the context of biological relevance of different types of techniques, we discuss about the strengths and limitations of the approaches and the need of prior botanical knowledge to reconstruct the plant geometry in biologically feasible manner. Finally, we explore the quantitative assessment techniques which can be used to measure the quality of the reconstruction result with respect to the actual data.