A semi-automatic protocol for delineating internal stem features from moisture-preserved discs using X-ray computed tomography

Accurate delineations of pith positions, tree-ring boundaries, and sapwood areas are essential for understanding tree growth dynamics, dendrochronology, and tree physiological mechanisms. While X-ray imaging has been exploited for wood specimens, few standardized and open-source methods are available. We present a supportive semi-automatic protocol, integrating free software, R and ImageJ, for delineating pith, tree rings, and sapwood areas of moisture-preserved stem discs with relatively high pixel resolution (150 µm). The protocol, documented on GitHub, integrates automated algorithms with interactive interfaces from ImageJ for manual corrections, providing a flexible and user-friendly framework. Pith position is determined by circularly fitting most-inner tree ring boundaries from edge-detection algorithms. This unique position is then used for cleaning edge detection results, supporting the two-dimensional grouping of tree ring segments. The Sapwood area is delineated by the Gaussian blur filter and interactive threshold interface. We tested the protocol on images of 15 moisture-preserved stem discs from three species (Picea abies (L.) H. KARST., Pinus sylvestris L., and Larix decidua MILL.). Our research demonstrated reliable accuracy, achieving an average mean error of 0.54 ± 0.42 mm for pith detection, 0.66 ± 1.72 mm for tree-ring boundaries, and 0.68 ± 1.23 mm for sapwood boundaries, offering support for manual delineations. While the method reliably supports analyses for coniferous species, challenges remain for ring-porous species, which require further advancements. The study serves as a milestone toward the future development of mobile X-ray systems for living trees in forest environments, offering a standardized and adaptable approach for advancing dendrology and wood science.