Automatic instantiation of a structural leaf model from 3D scanner data: application to light interception computation

A new parametric structural leaf model and its automatic instantiation from 3D digitized data is exposed. The proposed model allows representing the morphological features of a wide range of leaves into a controlled number of morphological parameters defining the accuracy of the 3D leaf representation. Using a set of suitable leaf-related geometrical hypothesis, these values are automatically inferred from dense digitized data provided by a non-contact digitizer Konica Minolta vi-910. A quantitative assessment, based on an inversion algorithm allowing building a tessellated leaf model from a prescribed set of parameters, demonstrates the reliability of both the model parameters and the instantiation process. An application to light interception computation via the software VegeSTAR (Adam & al. 2002) is also presented.