Study On Textile Color Fastness To Light Assessment Based On Imaging Processing Technology

The traditional method of textile color fastness to light assessment can't fit the modern need of the fabric appearance quality detection due to its high energy consumption,low reproducibility and so on.Our study set up a system insisted of hardware(high-resolution flat scanner and computer)and software(Color information science,imaging processing and expert system)to assess the color fastness to light.The main contents of investigation are the color calibration model,the method capturing the valid pixels on textile digital image,and the assessment model based on the blue references.In color calibration section,Colorchecker 140 card were utilized to get the digital color information RGB value and CIE LAB value,then the calibration model that presents the relationship between RGB and LAB was built through least square method.In method capturing the valid pixels on textile digital image section,the singular point was defined,then was deleted from the image,the pixels present the yarn gap was also deleted too.The RGB of a fabric was calculated based on the left pixels finally,and was transformed to CIE LAB value.In the model of textile color fastness to light assessment color assessment section,the CIE LAB values of blue reference was got through the way mentioned above,and calculate the CMC(2:1)and CIEDE2000 color difference of each level of blue reference.After comparing the outcome of two color difference expression,CMC(2:1)was more reasonable.The result of color difference was used to build assessment model through linear fit,fuzzy and neural network.The 282 samples which had assessed through traditional method was used to measure the system.The result showed that the digital method was a feasible way to replace the traditional method,but the totally substitution need more effort.