Advances in the industrial colour classification of artificial and natural textures based on new imaging colour appearance models

Abstract

This work is a first approximation to obtain an automatic classification system which solve the problem of over-classification at marble industries or other industries that have to do visual classification of its products. For this, we use some training samples to establish judgments for classifying colorimetrically and in an automatic way natural stone. Firstly, we measure the colorimetric values from the slab with a calibrated digital camera. After this, we implement the S-CIELAB model over the obtained images. Then we implement one spatial dithering method to simulate other viewing distances. Last, we calculate colour differences between the images and depending on its values we will establish judgments for classifying them.