Spectral imaging and a one-class classifier for detecting elastane in cotton fabrics.

Area of Science:

• Materials Science
• Analytical Chemistry
• Textile Engineering

Background:

• Elastane fibers in textiles pose challenges for mechanical and chemical recycling processes.
• Current methods for elastane detection are invasive and time-consuming, hindering efficient textile recycling.

Purpose of the Study:

• To develop a fast, non-invasive method for detecting elastane in cotton fabrics.
• To provide an alternative to existing time-consuming and invasive elastane detection techniques.

Main Methods:

• Utilized near-infrared (NIR) imaging spectroscopy combined with class modeling.
• Employed class-specific clustering for outlier identification and averaged pixel spectra to reduce uncertainty.
• Applied randomized resampling for robust classification accuracy assessment.

Main Results:

• Successfully detected 2-6% elastane in consumer cotton fabrics with high accuracy.
• Achieved median test set true positive and true negative rates of 89-97% through randomized resampling.
• Demonstrated the effectiveness of class modeling for identifying new material classes without exhaustive training data.

Conclusions:

• Near-infrared imaging spectroscopy and class modeling offer a viable solution for rapid, non-invasive elastane detection in textiles.
• This advancement moves closer to enabling a circular economy for textiles by improving recycling processes.
• The developed method provides a significant improvement over traditional, labor-intensive detection techniques.