ARTIFICIAL VISION INSPECTION APPLIED TO LEATHER QUALITY CONTROL

摘要

The paper deals with the problem of pointing out and mark the various kind of defects (mainly scars, wrinkles and fire brandings) of tanned calf leathers used in many industrial areas (e.g. household couches and car seats lining ones). In the first experiments here presented the authors try to use the same inspection method previously successfully applied to quality control of hands-made satin glasses, that is a two-dimensional wavelet-based de-noising technique of high resolution images of the leather under inspection; according to multiresolution analysis, this method produces a suitable number of decomposition levels of the image, then it carries out a thresholding operation on details and finally, using the threshold levels estimated considering the actual noise level, it assesses and mark the various kind of defects. The final aim of the work is the realization of an automatic and in-line computer vision system for leather analysis capable to accurately inspect and mark the defective areas of the analyzed leather specimens before they enter the production chain.