基于计算机视觉的织物光泽测试方法研究

Research on measuring method of fabric luster based on computer vision

织物光泽作为服装外观性能的特性之一,具有重要的研究意义。文章提出多角度织物光泽计算机视觉测试方法,通过数字图像处理技术提取光泽特征,运用随机森林回归模型并结合感官评价结果,进行织物光泽的测试与评价。该方法将计算机视觉应用于织物光泽的测试,可测试不同光照环境下的织物光泽。通过线性拟合分析,表明运用随机森林回归模型进行预测的结果,既符合织物光泽测试标准,也与感官评价结果相一致。实验结果表明,文章所提出的多角度织物光泽计算机视觉测试方法可有效用于织物光泽测试。

As one of the important characteristics of garment appearance,fabric luster is of great significance for study.The paper has proposed a multi-angle computer vision test method for fabric luster.The features of fabric luster were extracted through digital image processing technology,and the fabric luster was tested and evaluated by using Random Forest regression model combined with sensory evaluation results.This method has applied computer vision to fabric luster test,which can test fabric luster under different lighting environments.Through linear fitting analysis,it has been shown that the prediction results using the Random Forest regression model not only met the standard of fabric luster test,but also corresponded to the sensory evaluation results.The experimental results have shown that the multi-angle computer vision method proposed in this paper can be used effectively for the test of fabric luster.