Textile Sorting | AISORT
Application Overview — Textile Recycling
Textile Sorting in Modern Recycling Facilities
Textile sorting is transitioning from manual to automated as fiber-to-fiber recycling scales up. Automated textile sorting identifies fiber composition (cotton, polyester, wool, nylon, blends), color, and garment type — enabling high-purity feedstock for chemical and mechanical recycling processes. The EU's mandatory separate textile collection from 2025 is driving rapid investment in automated sorting capacity.
Material Characteristics and Sorting Challenges
Textile sorting challenges: fiber blends (polyester-cotton, wool-nylon) are difficult to classify with a single sensor; garment accessories (zippers, buttons, elastic, labels) are different materials than the main fabric; dyes and finishes alter the NIR signature of the base fiber; and textiles are flexible and irregularly shaped, making consistent sensor presentation difficult.
Recommended Sorting Technology Stack
NIR or hyperspectral (fiber composition identification: cotton vs polyester vs wool vs nylon vs acrylic vs blends) + RGB (color sorting into 50-100 color categories for fiber-to-fiber recycling) + AI vision (garment type recognition: t-shirt, denim, knitwear, outerwear). Conveyor-based systems with manual quality control stations for contaminated or ambiguous items.
Performance Benchmarks
| Metric | Target |
|---|---|
| Fiber ID Accuracy | >95% |
| Color Sorting | 50-100 categories |
| Throughput | 0.5-2 t/h per line |
| Garment Types | All post-consumer textiles |
These benchmarks represent achievable performance with modern sensor-based sorting equipment, assuming properly sized, well-maintained equipment operating on representative feedstock. Actual results depend on specific material composition, throughput, and operating conditions.