A computer vision approach for textile inspection

The economic viability of a weaving plant is significantly influenced by it s efficiency in eliminating fabric faults. Textile faults have traditionall y been detected by human visual inspection, but it is time consuming and do es not achieve a high degree of accuracy. Mechanisms with automatic operati on are much better. Industrial vision systems may form the basis of a metho d with a very high degree of accuracy for textile inspection. This work pre sents a simple system designed for fabric inspection and shows its efficien cy in detecting twelve kinds of common defects. The results show that the s uccess of system implementation depends on the choice of the specific appro ach.