您好，欢迎访问新疆畜牧科学院 机构知识库！

Rapid identification of scoured protein fibers using near-infrared spectroscopy with machine learning: A comparison of handheld and benchtop devices

文献类型： 外文期刊

作者： Gong, Ping 1 ; Feng, Yuchao 2 ; Wei, Peiling 1 ; Xu, Yanli 1 ; Zhang, Rongyin 1 ; Zheng, Wenxin 1 ; Fan, Xia 1 ;

作者机构： 1.Xinjiang Acad Anim Sci, Inst Anim Husb Qual Stand, Urumqi 830057, Peoples R China

2.Chinese Acad Agr Sci, Inst Qual Stand & Testing Technol Agroprod, Beijing 100081, Peoples R China

关键词： Cashmere; Wool; Convolutional neural networks; Handheld near-infrared spectrometer

期刊名称：MICROCHEMICAL JOURNAL （ 影响因子：5.1； 五年影响因子：4.7 ）

ISSN： 0026-265X

年卷期： 2025 年 208 卷

页码：

收录情况： SCI

摘要： Counterfeit cashmere textiles made from wool frequently appear in the market, disrupting regular commerce and significantly harming the interests of consumers. This study proposes the identification of scoured cashmere and wool fibers using a handheld near-infrared (NIR) spectroscopy system and then compared it with a benchtop Fourier transform-NIR spectroscopy instrument. A total of 416 fiber samples were analyzed, comprising 208 scoured cashmere and 208 scoured sheep-wool fibers. Prediction models were developed using multivariable selection algorithms, including principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), and one-dimensional convolutional neural network (1D-CNN). Competitive adaptive reweighted sampling (CARS) was employed to select characteristic wavelengths. Notably, the handheld NIR system demonstrated a predictive ability comparable to that of the benchtop NIR system, with 100% accuracy. Furthermore, the discriminant effect of PLS-DA was better than that of 1D-CNN, and the feature band could improve the performance of 1D-CNN model. For the study results contribute to the practical supervision and production of cashmere textiles in the market.

  • 相关文献
作者其他论文 更多>>
置顶
购物车
反馈

© 京ICP备09089781号-29 主办单位：新疆畜牧科学院

学术资源： 中国农业科学院 农业专业知识服务系统 农科联盟资源共建共享平台 中国农业科技文献与信息服务平台 中国农业期刊集成服务平台