Turmeric extracted from natural plants serves as a commonly natural dye but currently faces quality challenges due to the absence of standardization. Rapid determination of natural turmeric dye contents before the dyeing process is paramount. In this study, to determine the content of total and separate curcuminoid compounds, 155 samples of turmeric dyes were analyzed using both high-performance liquid chromatography and near-infrared technology. The near-infrared spectra within the range of 8000–5000 cm⁻¹ were selected, and after method optimization, the spectral preprocessing method of Savitzky-Golay smoothing (SG) + standard normal variate transformation (SNV) / multiplicative scatter correction (MSC) + first derivative (1st-Der) were used to construct the partial least squares (PLS) quantitative prediction models. Method validation results showed that the optimized model revealed exceptional prediction accuracy. In general, the total curcuminoid compounds quantitative prediction model demonstrated higher accuracy than that of the separate compounds, with R² > 0.99 and RPD > 10. In contrast, the separate curcuminoid compounds quantitative prediction model has R² > 0.97 and RPD > 6. Both models are suitable for turmeric dyes, and as fast and flexible detection methods, they are suitable for industrial production.

中文翻译：

---

使用近红外光谱法测定天然姜黄染料


从天然植物中提取的姜黄是一种常见的天然染料，但由于缺乏标准化，目前面临质量挑战。在染色过程之前快速测定天然姜黄染料的含量至关重要。在这项研究中，为了确定总姜黄素类化合物和分离姜黄素类化合物的含量，使用高效液相色谱和近红外技术分析了 155 个姜黄染料样品。选取 8000–5000 cm⁻¹ 范围内的近红外光谱，经过方法优化后，采用 Savitzky-Golay 平滑 （SG） + 标准正态变量变换 （SNV） / 乘法散射校正 （MSC） + 一阶导数 （1st-Der） 的光谱预处理方法构建偏最小二乘法 （PLS） 定量预测模型。方法验证结果表明，优化后的模型显示出优异的预测精度。总体而言，总姜黄素类化合物定量预测模型的准确性高于单独化合物，R² > 0.99，RPD > 10。相比之下，单独的姜黄素类化合物定量预测模型具有 R² > 0.97 和 RPD > 6。两种型号都适用于姜黄染料，作为快速灵活的检测方法，它们适用于工业生产。