Three-dimensional digital reconstruction techniques for ex-situ structural complexity-dependent ecological research

Abstract

Mangrove forests are integral components of tropical coastal ecosystems that provide numerous functions and services, which are derived from the three-dimensional structural complexity of the mangrove trees. However, experimental research regarding these functions have been limited by using simplistic replicas that bear little resemblance to true mangrove structures. Recently, improvements to 3D scanning technology have allowed quality reconstruction of complicated structures that makes realistic quantitative assessments possible. This study demonstrates the contrasting abilities of three technologies to scan mangrove structures of varying complexities at meter-scale. These tools, widely different in cost and usual application, utilize unique mechanisms for obtaining 3D objects: depth sensor in Microsoft Kinect, structure from motion photogrammetry, and lidar in GeoSLAM ZEB Revo RT. Each 3D scanning technique was able to reproduce digital models with relative accuracy and precision, albeit with individual strengths and limitations. Although Kinect remains as a reliable 3D scanner, the bulk of its accessories and sensitivity to sunlight hampered its use in routine field scanning. Photogrammetry reconstructed the most accurate mangrove tree structures, but was unable to reconstruct much of dense canopies and thin aerial roots that are characteristic of some species. ZEB Revo did not capture fine details essential in meter-scale mangrove tree structures, although its rapid and large-scale scanning capabilities make it valuable in mapping forest complexity and allows estimation of system attributes, e.g., blue carbon stock, at the forest scale. The potential for future applications of 3D printed mangrove models, especially in studying habitat use and animal behavior, was discussed.