Method for modeling plant root morphology based on ball splines skeleton

Plant growth modeling is very important in plant nutrition study. Based on the reconstructed 3D images of plant roots, the skeleton points and their radii are extracted by the Mini-crossing Section Algorithm (MCSA). Then the Many-knot Spline method is applied to interpolate a series of circles on the skeleton curve with some mesh points being sampled. The cross line of lateral and tap roots are also processed so that the mesh points of all roots could be connected smoothly. The geometric model that depicts the morphological traits is obtained by adding tiny roots and lighting the root surface. Our experimental results show that this modeling approach is very fast in computation and that the number of data points that determine the root morphology is several hundred times smaller than the original images. The root architecture model obtained by the algorithm is vivid. Compared to existing approaches, our algorithm is very robust on extracting skeleton points based on MCSA. As a result the surface morphology of roots is fitted efficiently to Ball Many-knot Spline.

Duke Scholars

Author Feng Tian DKU Faculty