Hybrid Approach for Precise Volume Estimation of Butressed Trees: Poisson Surface Reconstruction and TreeQSM Integration

Abstract

This study introduces a novel hybrid approach for accurately estimating the volume of trees with buttress formations, utilizing Terrestrial LiDAR Scanning (TLS) data. The hybrid method employs a semi-automated algorithm that first identifies buttress structures, divides trees into upper and lower portions (crown and buttressed trunk), and subsequently reconstructs these sections using TreeQSM and Poisson Surface Reconstruction (PSR), respectively. Our hybrid method demonstrates remarkable performance in volume estimation, surpassing TreeQSM with an average residual of 14.68%, compared to its 21.35%. The algorithm for buttress detection and segmentation achieves a high detection accuracy of 90.5% and an impressive 96.55% accuracy in precisely separating upper and lower portions. This fusion of PSR and TreeQSM holds substantial potential for advancing biomass estimation accuracy, optimizing forest management practices, and enhancing ecological research.