3D Visualisation of Specularity of Woven Fabrics

The research presents the workflow development for the creation of a specular map that serves for the visualisation of 3D specularity of a non-uniform, worn cloth surface. The aim of the experimental part was to analyse the influence of illumination settings during the image acquisition of the cloth data and the evaluation of the image processing techniques for a texture preparation. The analysed cloth was photographed at seven different illumination settings and the images were processed with two different image processing techniques, i.e. histogram equalisation and rolling ball algorithm. Including the original images, three groups of images were evaluated. To define the specular areas, the thresholding was performed with the Percentile algorithm and specular maps were evaluated with image analysis that was performed on specular and partially specular areas of interest (other parts were masked). The image analysis included the definition and evaluation of the area covered with specular surfaces, the average size of specular areas, the number of specular areas, the average value of pixel intensity, the most frequent value of pixel intensity and contrast. Further, specular maps were used in the cloth visualisations in 3D space at four different illumination directions. The specular areas of final visualisations after the rendering were analysed and the results were compared with the analysis of specular maps before the visualisation process. The findings confirmed the influence of the image processing technique on the image analysis results and on the distribution of specular areas in the specular map. Moreover, the influence of light direction and illumination settings during the image acquisition on optical phenomena on the cloth surface, and the relief effects of yarns and cloth (in the sense of distribution and size of specular area and pixel intensity of specular areas) are presented and discussed.