贮藏时间是影响菜心品质的一个重要因素，贮藏时间的常见鉴别方法主要依靠人工经验，这种方法准确度和可信度不高。该试验目标是基于拉曼光谱建立合适的化学计量学识别模型来对菜心贮藏时间进行判别。对菜心不同贮藏时间普遍对应的拉曼光谱进行评估，通过测试菜心采后每隔48 h的拉曼光谱，对预处理后的光谱数据进行主成分分析（Principal Component Analysis, PCA）、建立支持向量机（Support Vector Machine, SVM）模型和线性判别分析（Linear Discriminant Analysis, LDA）模型。结果表明，可溶性糖、纤维素类、可溶性蛋白和类胡萝卜素等是影响菜心贮藏品质变化的主要成分；PCA分析对菜心贮藏时间的鉴别整体的分界不明显，SVM模型下的线性和多项式函数分类准确率分别达97.75%和97.34%，验证集分类准确率达97.50%和96.00%。LDA二次函数模型分类准确率达99.00%，验证集准确率达97.00%。基于拉曼光谱建立的SVM模型和LDA模型均能有效识别菜心不同贮藏时间，为菜心以贮藏时间为新鲜度的识别提供参考和技术支持。

Storage time has a crucial effect on the quality of cabbage. Identifying the optimal storage time mainly relies on manual experience, which is not accurate and reliable. This study aimed to establish a stoichiometric recognition model based on Raman spectrum to calculate the storage time of cabbage. The Raman spectra corresponding to different storage times of cabbage heart were evaluated. Principal component analysis (PCA), support vector machine (SVM), and linear discriminant analysis (LDA) were established by testing the Raman spectra of cabbage heart every 48 h after harvest. The results showed that soluble sugars, cellulose, soluble proteins, and carotenoids were the main components affecting the storage quality of cabbage. The classification accuracy of linear and polynomial functions of the SVM model and verification set were 97.75 and 97.34%, respectively, and the classification accuracy of the SVM model and verification set were 97.50% and 96.00%, respectively. The classification accuracy of the LDA quadratic function model and the verification set were 99.00% and 97.00%, respectively. Both SVM and LDA models based on Raman spectrum effectively identified different optimal storage times of cabbage. Thus, this study provides reference and technical support for the identification of cabbage freshness based on storage time.

广东省科技厅农村科技特派员项目（KTP20190079）；仲恺农业工程学院-广州酒家产学研合作项目（D11820760）