Aggregation as similarity in a morphological framework for the processing of textile images

Mathematical morphology has demonstrated to be a very powerful and flexible mean for the processing of binary and grayvalue images heretofore. Nevertheless, the extension of mathematical morphology to color images results are non-trivial due to the vectorial nature of such signals. This work presents an approach on color morphology based on the fuzzy integral, which is employed within the framework for the computation of similarities through the application of fuzzy aggregation operators. The main feature of the new defined morphological framework is the possibility to apply the so-called targeted morphological operations on the treated images. The usefulness of this approach is demonstrates the results of different pre-processing operations for an automated visual inspection system, whereby textile images are inspected.