Color calibration for fabric image analysis based on spectral reflectance reconstruction

Abstract

Spectral characterization can be used for the color calibration of the fabric image to eliminate the influence of the imaging equipment and illumination environment, so as to obtain the actual yarn color information of the fabric image and improve the accuracy of fabric color analysis and rendering. However, existing algorithms based on spectral reflectance reconstruction were not ideal to minimize the obvious c between the reconstructed and its actual color. Compared with the index of spectral reflectance reconstruction accuracy, the chromaticity error is a direct index of image color quality evaluation by human vision. Therefore, to eliminate the influence of imaging equipment and illumination environment on the fabric image color occurring during the fabric image acquisition, a novel color calibration method based on spectral reflectance reconstruction was proposed for fabric image analysis in this paper, which focused on the optimal selection of representative colors and defined a new spectral reconstruction error evaluation function E=αErmse+βECA. Experimental results show that the proposed method in this paper is superior to the existing methods in color reconstruction accuracy and visual quality.