Abstract
Shipworm (Teredo navalis) poses a significant threat to wooden structures and archaeological artifacts in marine
environments, leading to extensive biodeterioration. Quantifying shipworm damage remains challenging due to
the limitations of traditional assessment methods, which are often invasive, time-consuming, and/or subjective.
In this study, we introduce RANDA (Rapid Analysis Digital Tool), a software designed to assess shipworm
damage severity through image-based analysis of entry holes on wood surfaces, with the potential for nondestructive
application. Data from 232 T. navalis entry holes and tunnels revealed a strong correlation between
entry hole size and tunnel volume, suggesting that surface analysis can approximate internal damage.
When applied to eight test panels, RANDA degradation estimates differed by -0.69±1.05 % from weight loss
analysis and by 2.74 ± 1.67 % from tunnel volume measurements obtained via CT imaging. For hole classification,
the RANDA model achieved a precision of 80 % and a recall of 81 %. While currently validated under
controlled conditions, this method offers a rapid and reliable alternative to conventional approaches and represents
an important step toward a non-destructive in situ assessment tool.
environments, leading to extensive biodeterioration. Quantifying shipworm damage remains challenging due to
the limitations of traditional assessment methods, which are often invasive, time-consuming, and/or subjective.
In this study, we introduce RANDA (Rapid Analysis Digital Tool), a software designed to assess shipworm
damage severity through image-based analysis of entry holes on wood surfaces, with the potential for nondestructive
application. Data from 232 T. navalis entry holes and tunnels revealed a strong correlation between
entry hole size and tunnel volume, suggesting that surface analysis can approximate internal damage.
When applied to eight test panels, RANDA degradation estimates differed by -0.69±1.05 % from weight loss
analysis and by 2.74 ± 1.67 % from tunnel volume measurements obtained via CT imaging. For hole classification,
the RANDA model achieved a precision of 80 % and a recall of 81 %. While currently validated under
controlled conditions, this method offers a rapid and reliable alternative to conventional approaches and represents
an important step toward a non-destructive in situ assessment tool.
| Originalsprog | Engelsk |
|---|---|
| Artikelnummer | 105988 |
| Tidsskrift | Results in Engineering |
| Vol/bind | 27 |
| Udgave nummer | 2025 |
| Sider (fra-til) | 1-10 |
| Antal sider | 10 |
| ISSN | 2590-1230 |
| DOI | |
| Status | Udgivet - 25 jun. 2025 |
Kunstnerisk udviklingsvirksomhed (KUV)
- Nej
Citationsformater
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver