New Methods in Digital Wood Anatomy: The Use of Pixel-Contrast Densitometry with Example of Angiosperm Shrubs in Southern Siberia.

Abstract

This methodological study describes the adaptation of a new method in digital wood anatomy, pixel-contrast densitometry, for angiosperm species. The new method was tested on eight species of shrubs and small trees in Southern Siberia, whose wood structure varies from ring-porous to diffuse-porous, with different spatial organizations of vessels. A two-step transformation of wood cross-section photographs by smoothing and Otsu's classification algorithm was proposed to separate images into cell wall areas and empty spaces within (lumen) and between cells. Good synchronicity between measurements within the ring allowed us to create profiles of wood porosity (proportion of empty spaces) describing the growth ring structure and capturing inter-annual differences between rings. For longer-lived species, 14-32-year series from at least ten specimens were measured. Their analysis revealed that maximum (for all wood types), mean, and minimum porosity (for diffuse-porous wood) in the ring have common external signals, mostly independent of ring width, i.e., they can be used as ecological indicators. Further research directions include a comparison of this method with other approaches in densitometry, clarification of sample processing, and the extraction of ecologically meaningful data from wood structures.

Plant phenotyping relevance

樹木断面画像から木材孔隙率・年輪構造を抽出する画像解析法を開発・検証しており、植物形質の取得方法が研究の中心です。

Code and data availability

The paper's authors publicly released the Python software implementing their PiC densitometry method on GitHub (OpenPiCDens), and the MDPI supplement contains paper-specific wood cross-section photographs, binary images, and porosity profiles. The underlying raw measurement data are only available on request.