Spatial Consistency Improvement And Yarn Color Measurement In Multispectral Imaging System

As the development of color measurement technology,people have a higher demand for accurate measurement of colors under various applications.Traditional method,which measures spectral reflectance using a spectrophotometer,cannot be applied to non-solid colors and thus has limitations in industrial applications.Multispectral imaging system can get not only the spectrum but also the image of the sample,so it is becoming more and more important in the field of color measurement.The research work in this thesis is based on the multispectral imaging color measurement(ICM)system developed in our laboratory.We propose a spatial consistency correction algorithm to improve the spatial consistency of the ICM system.We also propose a yarm color measurement method based on ICM system to solve the color measurement issue of short yarn.There exists a spatial consistency problem in the ICM system.Although we have used the standard whiteboard to correct spatial light intensity,the measurement results of a sample still vary at the different positions of sample holder.This problem is caused by the different reflection characteristics between the surfaces of textile fabric and standard whiteboard.To solve this problem,this thesis proposes a spatial consistency correction method based on polynomial regression,and compares it with the virtual-board method.The experimental results show that the proposed method not only outperforms the virtual-board method on improving spatial consistency,but also increases the color measurement accuracy of the system.Conventional yarm color measurement method usually measures the spectral reflectance of the corresponding yarn board,but it cannot work when the yarn is too short to wind on the yarn board.To solve this problem,a yarn color measurement method is proposed in this thesis.The method introduces a special yarn winding device and uses a kernel ridge regression model to map the reflectance measured by the ICM system to the reflectance measured by standard spectrophotometer.We compare the proposed method with traditional linear regression,ridge regression and local linear regression.The experimental results show that,the proposed method performs better than traditional methods,and well solves the color measurement issue of short yarns.