染整技术2026，Vol.48Issue(2)：12-19,55,9.
基于机器学习的纺织品洗液沾色程度AI评估模型构建
Construction of AI evaluation model for color transfer degree of textile washing liquid based on machine learning
摘要
Abstract
To address the challenges in emerging textile washing solution staining evaluation,such as the lack of standardized methods,high subjectivity,low efficiency,and unquantifiable differences caused by reliance on visual inspec-tion,a feature optimization strategy that integrates mutual information,Pearson correlation coefficient,and multi-algorithm joint screening is proposed.By comparing six machine learning models with dE(color difference)as the key feature,the K-nearest neighbors(KNN)model achieves an accuracy of 92%.This model not only provides a reliable solution for assessing the degree of textile washing solution staining,but also enables the transition from conventional visual-based evaluation to automated model-driven grading,significantly enhancing objectivity and operational efficiency in industrial applications.关键词
分光光度技术/色差/机器学习/K近邻Key words
spectrophotometric technology/color difference/machine learning/K-nearest neighbors分类
轻工纺织引用本文复制引用
陈威,钱琴芳,胡一飞,沈俊..基于机器学习的纺织品洗液沾色程度AI评估模型构建[J].染整技术,2026,48(2):12-19,55,9.基金项目
海关科研项目(2023WK007) （2023WK007）
