LINEAR FLAW DETECTION IN WOVEN TEXTILES USING MODEL-BASED CLUSTERING

We combine image-processing techniques with a powerful new statistical technique to detect linear pattern production faults in woven textile s. Our approach detects a linear pattern in preprocessed images via mo del-based clustering. It employs an approximate Bayes factor which pro vides a criterion for assessing the evidence for the presence of a def ect. The model used in experimentation is a (possibly highly elliptica l) Gaussian cloud superimposed on Poisson clutter. Results are shown f or some representative examples, and contrasted with a Hough transform . Software for the statistical modeling is available. (C) 1997 Elsevie r Science B.V.