Development of a machine vision system for on-loom and real-time defect detection of woven fabrics with different constructions

The purpose of this study is developing the machine vision system on the weaving machine and detection of fabric defect during the weaving process. It was aimed that the defective regions on the fabric surface will be detected and occurrence of these defects among long meters will be prevented.Weft lacking, oil stain and fly type defects were determined for detection. The algorithm application studies were conducted over totally 200 image frames by acquiring 50 defect sample image frames for every defect type. The selected algorithm was applied real time during the weaving process. Use interface was prepared for real-time application. It was revealed that the developed system and image processing algorithm can be applied on an industrial weaving loom.