Research On Textile Color Characterization And Color Fastness Grading Method Based On Hyperspectral Imaging Technology

The resistance of textile colors to various types of manipulation during processing or using is called color fastness.Color fastness grading is an extremely important indicator for controlling the quality of textile production and testing the quality of finished textile products.These days,color fastness grading evaluation mostly follows the method of the traditional gaze rating,but the result of the gaze rating is susceptible to the influence of external environmental factors and subjective factors of the rating staff.Moreover,there are many types of quality inspection items about the textile color fastness and a large amount of inspection,which bring a heavy workload to the staff of textile quality inspection.At present,many researchers have tried to use different kinds of instruments for making color fastness grading.This type of method has the advantages of objectivity and stability,which can solve the problem of strong subjectivity produced by the gaze rating.However,the existing instruments also have some drawbacks,such as only a smaller area on the fabric can be sampled by these instruments,and the range of colors that can be recorded and reproduced is limited.These defects lead to the fact that the evaluation method of the instrument-based color fastness has not been widely used and can only be used as a reference rather than be used as a substitute for the traditional gaze rating method.Therefore,developing and establishing a both fast and accurate textile color fastness grading method is an urgent need.This project studies on the textile color fastness grading method based on hyperspectral imaging technology,which realizes the innovative integration of spectral analysis and traditional textile test.These are the primary research contents:the acquisition conditions of applicable hyperspectral image are analyzed for textiles.And on this basis,the spectral reflectance curves of specimens and their original samples after different color fastness testing experiments are obtained.The spectral similarity is used to measure the difference between the spectra of each grade of the standard discolored gray sample cards,and the relationship equation between each grade of the standard discolored gray sample cards and the spectral similarity is obtained by means of fitting,so as to establishing a color fastness grading model based on spectral similarity.The results of this research propose a color fastness grading model based on hyperspectral imaging technology,which is expected to replace the manual gaze grading.It makes up for the shortcomings of the current instrumentation method,expands the application prospects of hyperspectral imaging technology in the field of textile quality inspection,and provides new ideas for the further combination and application of hyperspectral imaging technology and textile color fastness inspection in multiple fields.The specific conclusions are as follows:(1)Comparing the effects of different acquisition conditions on hyperspectral imaging and spectral data,changes in focal length and speed conditions will ha-ve a certain effect on hyperspectral imaging,but almost have no effect on the e-xtracted spectral data.The effect of external lighting conditions on the original s-pectrum can be eliminated by black and white plate correction.Single-layer fab-ric will not affect the subsequent grading model under the same acquisition con-ditions.The black placement plate has the least effect on the spectral curve of s-inglelayer fabric and is most suitable as a background plate.(2)Characterization of the spectral curves of the specimen and its original sample after different testing experiments on the color fastness shows that the color fastness testing experiments have a certain influence on the reflectance value.The spectral curves of the tested and original samples have deviated to a certain extent.And as the color difference increasing,the distance between the separated parts of the curves increases accordingly.The color fastness grading can be distinguished according to the deviation degree of the spectral curves before and after color fastness testing.(3)Combining the relationship between spectral similarity and discoloration grade among the grades of discoloration gray sample cards,the spectral similarit-y-discoloration grading equation can be obtained,and the best discoloration color fastness grading equationD(28)4.946-9.051X_c _d(10)5.382X_c~2 _d(X_c _d is the calculated spectral similarity value)is optimized by using the manual gaze rating results.T-he color fastness grading model based on spectral similarity is established by us-ing this equation.(4)The actual grading results of the discoloration color fastness grades of 200 sets of samples were verified to compare the differences in results between the grading model and the gaze ratings.The results showed that the grading model established in this project was 91%accurate within the half-grade difference allowed for the color fastness grade results,and there was no significant difference between the two.Compared with the more subjective gaze rating,the grading model established in this project has certain advantages.