Computer vision in garment and textile manufacturing mainly covers quality control, i.e. maintaining premium quality throughout all production stages. Errors and anomalies can be discovered in various fabric patterns such as weaving, knitting, braiding, finishing and printing. The automated detection of structural and surface defects, based on color and texture variations, allows for driving feedback loops to guide production decisions. Detecting issues earlier in the process enables manufacturers to further reduce cost repercussions as well as the amount of waste.
In many applications, a line camera is used in combination with AI driven data processing, possibly including deep learning. The images-based processing enables to assess the quality of work, while the inspected textile material is moving at high speed across a production line. Historical and real-time operational data are essential in leveraging AI to improve production quality and efficiency. Typical inspection applications that can be further improved and automated include textile pattern inspection, color matching and fabric wrinkle detection.
Textile pattern inspection
Finding defects on textiles with various color and texture-based patterns is challenging. Computer vision and AI driven data processing systems can be trained to reliably distinguish ink stains or stitching errors from regular fabric patterns. By teaching tens or hundreds of 'good' and 'bad' textile samples, such systems learn fabric patterns, yarn characteristics and color distributions. After a period of training, the systems are able to reliably detect defects in textile end products.
Automatic fabric wrinkle detection
Wrinkles in fabric materials should be taken into account during inspection because they influence the way people will perceive garments in shops and boutiques. Computer vision and AI driven data processing focused on the size, orientation and distribution of fabric wrinkles potentially enhance decision making sophistication. The automated judgement introduces higher inspection productivity and delivers more consistent fabric quality.
