Rijken, Temmo (2026) Machine Learning–Based Identification of Textile Composition from NIR Spectroscopy for Automated Sorting. Research Project, Industrial Engineering and Management.
|
Text
MScReportResearchProjectTemmoRijken-Machine-LearningBased-Identification-of-Textile-Composition-from-NIR-Spectroscopy-for-Automated-Sorting.pdf Download (4MB) | Preview |
Abstract
The growing volume and complexity of post- consumer textile waste make accurate material identification increasingly important for automated sorting and high- value recycling. This study investigates whether near-infrared (NIR) spectroscopy combined with machine learning can classify recycling-relevant textile compositions and detect elastane. A dataset of 181 textile samples was collected from certified suppliers and measured in the 900–1700 nm range using a benchtop NIR spectrophotometer. Spectral, ratio-based, and wavelet-based features were extracted, after which feature selection and supervised classification were applied for composition identification, while a 1D convo- lutional neural network was developed for binary elastane detection. The best composition classification model, based on Weighted KNN, achieved an accuracy of 91.7%, showing that common mono-materials and blends can be distinguished reliably. The elastane detection model achieved an overall accuracy of 92.6%, with class-specific accuracies of 83.3% for elastane-present samples and 95.2% for elastane-absent samples. These results show that NIR spectroscopy combined with machine learning is a promising approach for automated textile sorting and for improving material recovery in circular textile recycling.
| Item Type: | Thesis (Research Project) |
|---|---|
| Supervisor name: | Munoz Arias, M. and Cheng, L. |
| Degree programme: | Industrial Engineering and Management |
| Thesis type: | Research Project |
| Language: | English |
| Date Deposited: | 30 Apr 2026 08:43 |
| Last Modified: | 30 Apr 2026 08:43 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/37298 |
Actions (login required)
 |
View Item |