CS-TRD: a Cross Sections Tree Ring Detection method

TL;DR

This paper introduces CS-TRD, a fully automated method for detecting and linking tree rings on complete cross-sections. It leverages a pith-centered spider-web sampling framework and an edge-based approach using the Canny-Devernay detector, followed by iterative chain sampling, connectivity, and postprocessing to form closed ring contours. Evaluations on the UruDendro (Pinus taeda) and Kennel (Abies alba) datasets show strong performance with $F$-Scores of $0.89$ and $0.97$, respectively, and competitive runtimes on CPU without specialized hardware. Compared to the INBD method, CS-TRD achieves higher precision and recall, demonstrating robustness to perturbations like knots and fungal stains. The work suggests future directions including automatic pith detection and cross-species extension, with potential speedups via optimized implementations.

Abstract

This work describes a Tree Ring Detection method for complete Cross-Sections of Trees (CS-TRD) that detects, processes and connects edges corresponding to the tree's growth rings. The method depends on the parameters for the Canny Devernay edge detector (sigma), a resize factor, the number of rays, and the pith location. The first five are fixed by default. The pith location can be marked manually or using an automatic pith detection algorithm. Besides the pith localization, CS-TRD is fully automated and achieves an F-Score of 89% in the UruDendro dataset (of Pinus taeda) and 97% in the Kennel dataset (of Abies alba) without specialized hardware requirements.

Figures (18)

• Figure 1: Examples of core tree-ring images taken from a dataset with 239 images FABIJANSKA2017279.
• Figure 2: Some examples of images from the UruDendro dataset. Note the variability of the images and the presence of fungus (image L02b), knots (images F07b and F03c), and cracks (images F02e and L03c). The first five images are from the same tree at different heights, as the text explains in Section \ref{['sec:database']}.
• Figure 3: (a) The whole structure, called spider web, is formed by a center (which corresponds to the slice pith), $N_r$rays (in the drawing $N_r=18$) and the rings (concentric curves). In the scheme, the rings are circles, but in practice, they can be (strongly) deformed as long as they don't intersect another ring. Each ray intersects a ring only once in a point called node. (b) A curve is a set of connected points (small green dots). Some of those points are the intersection with rays, named nodes (black dots). A chain is a set of connected nodes. In this case, the node$N_i$ is the point$p_n$. (c) Each Chain$Ch_k$ and $Ch_{k+1}$, intersects the rays$r_{m-1}$, $r_{m}$ and $r_{m+1}$ in nodes$N_{i-1}$, $N_{i}$ and $N_{i+1}$.
• Figure 4: Principal steps of the CS-TRD algorithm: (a) Original image, (b) Background subtraction, (c) Pre-processed image (resized, equalized, grayscale conversion), (d) Canny Devernay edge detector, (e) Edges filtered by the direction of the gradient, (f) Detected chains, (g) Connected chains, (h) Post-processed chains and (i) Detected tree-rings.
• Figure 5: A given chain (in black) with two endpoints A and B. Its nodes (in red) appear at the intersection between the Canny Devernay curve and the rays. The ray at endpoint A is in blue. Other chains detected by Canny Devernay are in white. Endpoint A's inward and outward chains are in yellow and orange, respectively.