基于图像光影重构的缎纹影光织物明度预测方法

Lightness prediction method for shaded satin fabrics based on image reconstruction of light and shadow

为更准确获取缎纹影光组织织物的明度值,提出一种基于织物图像光影重构的明度预测方法。将织物图像分离为经纬图形层、阴影层和材质层,计算3层中不同因素的相对面积率,并赋予各图形面积以实际经纬纱线明度值,采用明度相加重构织物图像明度值,再将重构明度与织物实际测色数据进行拟合分析。选择经纬图形、投影和杂点面积率为自变量,建立经、纬两向加强织物的明度预测回归模型,并随机选取样本检验模型精确度。结果表明:经纬图形层重构明度与织物实测明度的接近度为0.15,依次增加阴影层和材质层明度,则接近度从0.76提升为0.89;模型预测的织物明度总体相对误差在±4%以内,预测结果可为影光组织织物设计提供参考。

In order to obtain the brightness value of shaded satin fabrics more accurately,a lightness prediction method based on the shadow reconstruction of fabric image was proposed.The fabric image was separated into three layers:warp and weft graphic layer,shadow layer and texture layer,and the actual color values of the warp and the weft yarn were assigned to the relative area of the three separated image layers when their pixel numbers were counted respectively.After the fabric image brightness value were reconstructed by adding the lightness values,fitting analysis were implemented between the reconstructed one and the actual color measurement data of the fabric.According to the analysis results,the warp and weft area rate,shadow area rate and texture area rate were selected as the independent variables.A regression model for the lightness prediction of the fabric was established based on the three independent variables,and random samples were selected to test the accuracy of the model.The results show that the proximity between the lightness of the warp and weft graphic layer and the measured lightness is 0.15,and the proximity increases from 0.76 to 0.89 when the shadow layer and material layer are added in sequence.The overall relative error of the fabric lightness prediction by the model is within±4%,the prediction results can be used as references for the design of the fabric with shaded weaves.