A novel set of algorithms to recognize galleries of ambrosia beetle in computerized axial tomography of trees trunks

TL;DR

This paper introduces a set of algorithms for automatically recognizing and isolating galleries of ambrosia beetles in CT images of tree trunks, enabling non-destructive analysis of insect galleries.

Contribution

The paper presents a novel, robust algorithm for automated extraction of beetle galleries from CT scans, improving non-destructive analysis of insect structures within trees.

Findings

Effective gallery isolation from CT images achieved

Algorithm handles low contrast between galleries and background

Enables detailed 3D analysis of insect galleries

Abstract

Megaplatypus mutatus is an ambrosia beetle that attacks several species of trees by making galleries in the trunks where its larvae and associated fungi develop. This damage spoils the wood for commercial use and cause stem breakage in front of strong winds. Due to the insect's cryptic lifestyle, gallery analyses have usually been studied by destructive methods. However, they alter the homeostasis of the insect-fungi interaction, modifies the topology of the gallery and, more importantly, does not reveal the 100% complex structure made by the insect. Therefore, a novel way to study this structure is by imaging the galleries by means of computerized axial tomography (CT). This method allows obtaining a three-dimensional representation of the gallery and the pupal chambers to be studied, while the wood and insect sample is not disturbed and generates a high amount of data. The isolation of the galleries and pupal chambers from the CT background images is not simple, because there is not enough contrast between the grey levels of the galleries and the marks generated by internal components of the trunk itself. In this paper, we present a robust algorithm that allows automating the isolation of the galleries and pupal chambers from CT trunk images which can be used in a broad spectrum of image analysis.