| Title |
Image analysis methods for assessing fibrous microplastic release |
| Authors |
Gliaudelyte, Ugne |
| DOI |
10.25368/2025.006 |
| Full Text |
|
| Is Part of |
AUTEX 2025 World conference, 11-13 June 2025, Dresden, Germanyc: book of abstracts / editors: O. Kyzymchuk, Y. Kyosev, C. Cherif.. Dresden : Technische Universit¨at Dresden. 2025, p. 29 |
| Keywords [eng] |
microfiber ; microplastic ; textile ; sustainability |
| Abstract [eng] |
Fibrous microplastics can be released into the environment through fiber breakage due to various factors. Analysis using light microscopes and digital cameras is among the most popular methods for studying fibrous microplastics due to its accessibility. However, manually processing images and assessing fibrous microplastics is highly time-consuming. This study compares different analytical image analysis methods to find the most suitable option for effective fibrous microplastics assessment. To analyze the fibrous microplastics, three types of image analysis methods were applied: manual analysis with ImageJ, semi-automatic analysis with ImageJ’s Analyze Particles plugin, and automatic analysis using AI-based Ilastik software. For the evaluation, a black knitted fabric was used with the following specifications: composition – 95% polyester / 5% elastane; mass per unit area – 224.0 g/m2; thickness – 0.52 mm; course density – 23.0±0.5 cm−1; wale density – 22.0 ± 0.5 cm−1. Three specimens were prepared. Individual washing tests were performed on each specimen. Wastewater was filtrated through filters after each wash. Each filter was photographed using light microscope (Lumenera Infinity) and digital camera (Nikon D5600) with lens (AF-P NIKKOR 18-55mm f/3.5- 5.6G). Results indicated that manual, semi-automatic and automatic methods were similarly effective in assessing non-overlapping microplastics. Images taken with the light microscope provided more accurate results. After training the AI-based software, the automatic method was the least time consuming and manual analysis took the longest. However, neither semi-automatic nor automatic methods were performing effectively on overlapping fibrous microplastics and manual corrections were required to attain more accurate results. These findings suggest that AI-based automatic software can be the most effective and the least time-consuming solution for analyzing fibrous microplastics when they are wellseparated on surfaces, but it needs further development to improve the analysis for overlapping fibrous microplastics. |
| Published |
Dresden : Technische Universit¨at Dresden |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2025 |
| CC license |
|
