The Turkish knit fabric producer shows how AI inspection can move quality control from end-stage rejection to real-time defect prevention.
Ekoten Tekstil, a major weft-knit fabric manufacturer near İzmir, has implemented an AI-driven quality-control system that has sharply reduced knitting defects while improving productivity, energy use and customer responsiveness. The company serves fashion, performance workwear, military, technical, activewear and sportswear markets, supported by 235 knitting machines and daily knitting capacity of about 40 tonnes.
Scale makes inspection critical
Ekoten’s production base includes jet and pad dyeing, digital printing, 11 stenters and a broad finishing line. Its in-house colour laboratory can process up to 700 lab dips per day, while the company develops around 250 new fabric designs each month. At this level of product variety, manual inspection alone becomes difficult to scale because defect patterns, fabric constructions and colour requirements change rapidly.
From defect detection to prevention
The company began researching AI-based knitting defect prevention in 2019 and installed real-time inspection systems on 29 knitting machines. A one-year assessment in 2025 showed that the system prevented 2,454 hours of defective production, avoided 32,100 kg of faulty fabric, saved nearly 289,000 kWh of electricity and cut CO₂ emissions by 71,936 kg.
Ekoten has since extended AI quality control into dyeing and finishing. After studying inspection points for six months, the company moved monitoring to the stenter exit, allowing simultaneous detection of surface faults and dyeing inconsistencies. The system combines AI cameras, real-time monitoring software, automated reporting and in-line spectrophotometer-based colour measurement. Defects are classified, mapped by precise coordinates and reported instantly.
Faster inspection, less variation
Compared with manual inspection, Ekoten’s AI system has increased inspection speed from 8–15 metres per minute to 30–50 metres per minute, with reported accuracy of 95%. Productivity has risen to as much as 2,500 metres per hour, while operator variability and training time for new fabrics have been reduced.
The wider lesson for knitters is clear: AI inspection is becoming an operational tool, not a showroom technology. The next frontier will be integrating defect data with machine settings, maintenance planning and customer quality feedback to prevent faults before fabric is produced.
