目的 分析苦杏仁Armeniacae Semen Amarum炒制过程中色泽、气味的动态变化规律，建立炒苦杏仁“火候”判别模型，为其智能化生产和应用提供科学依据。方法 制备不同炒制程度的苦杏仁样品，利用测色仪和电子鼻技术获取各苦杏仁样品色泽、气味数据信息，结合层次聚类分析（hierarchical clustering analysis，HCA）、偏最小二乘-判别分析（partial least squares-discriminant analysis，PLS-DA）、秩相关分析（rank correlation analysis，RCA）和线性判别分析（linear discriminant analysis，LDA）等方法构建判别模型。结果 随着炒制程度的加深，苦杏仁的明度（L^*）逐渐降低，红绿色度（a^*）和黄蓝色度（b^*）显著增加；气味方面，生品在W1S传感器响应值最强，而炒制品在W5S、W1W、W2W传感器上的响应值随炒制程度加深而增大。基于色泽参数L^*、a^*、b^*及气味参数建立的判别模型，对苦杏仁生品及不同程度炒制品的判别正确率分别达到98%和96%。通过百分位数法进一步计算了不同炒制程度苦杏仁的色泽和气味双侧90%参考范围。结论 测色仪和电子鼻技术可快速、准确地辨识苦杏仁的炒制程度，为苦杏仁的质量控制提供了科学依据，同时为其他种子类中药的炒制程度快速辨识研究提供了新思路。

Objective This study aims to investigate the dynamic changes in color and odor during the stir-frying process of Kuxingren (Armeniacae Semen Amarum, ASA), and to establish a “fire control” discriminant model for stir-fried ASA, providing a scientific basis for intelligent production and application.Methods ASA samples with different degrees of stir-frying were prepared, and their color (L^*, a^*, b^*) and odor data were obtained using a colorimeter and electronic nose technology. Discriminant models were constructed using hierarchical clustering analysis (HCA), partial least squares-discriminant analysis (PLS-DA), rank correlation analysis (RCA) and linear discriminant analysis (LDA).Results The results indicated that as the degree of stir-frying increased, the lightness (L^*) of ASA decreased. While the redness (a^*) and yellowness (b^*) increased significantly. In terms of odor, the raw samples exhibited the highest response value on the W1S sensor, while the stir-fried samples showed increased response values on the W5S, W1W, and W2W sensors with increasing stir-frying degree. The discrimination models based on color parameters (L^*, a^*, b^*) and odor parameters achieved discriminant accuracies of 98% and 96%, respectively. The bilateral 90% reference ranges for color and odor of ASA with different stir-frying degrees were further calculated using the percentile method.Conclusion This study demonstrates that colorimeter and electronic nose technology can rapidly and accurately identify the stir-frying degree of ASA. providing a scientific basis for quality control of ASA, and offering a new approach for the study of processing degree identification of other traditional Chinese medicinal materials.

R283.6

国家重点研发计划（2023YFC3504200）；全国名老中医药专家传承工作室建设项目（国中医药人教函〔2022〕75号）