Constant and traceable high fabric quality is of high importance both for technical and for high-quality conventional
fabrics. Usually, quality inspection is carried out by trained personal, whose detection rate and maximum period of
concentration are limited. Low resolution automated fabric inspection machines using texture analysis were developed.
Since 2003, systems for the in-process inspection on weaving machines ("onloom") are commercially available. With
these defects can be detected, but not measured quantitative precisely. Most systems are also prone to inevitable machine
vibrations. Feedback loops for fault prevention are not established.
Technology has evolved since 2003: Camera and computer prices dropped, resolutions were enhanced, recording speeds
increased. These are the preconditions for real-time processing of high-resolution images. So far, these new
technological achievements are not used in textile fabric production. For efficient use, a measurement system must be
integrated into the weaving process; new algorithms for defect detection and measurement must be developed.
The goal of the joint project is the development of a modern machine vision system for nondestructive onloom fabric
inspection. The system consists of a vibration-resistant machine integration, a high-resolution machine vision system,
and new, reliable, and robust algorithms with quality database for defect documentation. The system is meant to detect,
measure, and classify at least 80 % of economically relevant defects. Concepts for feedback loops into the weaving
process will be pointed out.