Colorful Math: Developing Algorithmic Methodology to Visualize and Analyze the Dynamics of a Deciduous Tree

Abstract

The purpose of this study is to determine if it is possible to use LIDAR (light detection and ranging) point cloud data collected using a FARO Focus3D laser scanner to: 1) generate a high quality 3D virtual representation of a scanned tree and reduce noise due to environmental variability, 2) develop an algorithm to differentiate between different sections of trees based on RGB values associated with these point clouds and calculate the respective volumes of these sections, and 3) project leaf color change dynamics over time. The visualization tools used are Matlab and R. Additional analysis was done with a high performance computing cluster (Clemson University's Palmetto cluster). The data is of an American Sycamore Platanus occidentalis over a period of eight weeks as the tree transitions from fall into winter. The conclusions of this study have implications in outlining innovative approaches to three-dimensional tree dynamic measurement using advanced technology and tools. These approaches can provide the baseline information of a forest structure for greater understanding of how a tree changes seasonally, as well as the establishment of methodology for making laser scanning software and Matlab's Image Processing Toolbox more flexible to use. The visualization aids in this study can inform and be used to estimate with greater precision the effects that disturbances, such as fires and hurricanes, have on forest structures.