The textile industry is highly competitive, with decreasing production costs and increasing disposal of clothes. Currently, such items are mostly combusted, disposed, or downcycled into lower value products like fleece, posing an ever-increasing demand on (fossil) resources and energy. Only a small share of garments is reused, as the sorting process is traditionally performed by trained personnel, resulting in significant time consumption. The method presented here employs computer vision and artificial intelligence for textile pre-sorting, significantly enhancing the proportion of reusable textiles. The method relies on a custom high-resolution camera system that captures images of textiles and garments on a conveyor belt. The system not only recognizes the type but also infers the quality as well as the fabric type of garments, ensuring high-quality sorting and substantially increasing the proportion of textiles that can be reused. The presented approach of fabric detection in visually inspected objects has potential applications beyond post-consumer sorting, such as in electronics and plastics recycling. It can also be applied to manufacturing tasks across various industries—including agriculture, food, and automotive—enhancing quality control, reducing waste, increasing efficiency, and thus supporting the principles of a circular economy.