Spectra and image features combined with machine learning for drying temperature identification of Phlebopus portentosus

文献类型： 外文期刊

第一作者： Ji, Zhiyi

作者： Ji, Zhiyi;Li, Jieqing;Ji, Zhiyi;Wang, Yuanzhong;Liu, Honggao

作者机构：

关键词： Infrared spectroscopy; Drying temperature; Image features; Data fusion; Deep learning

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 212 卷

页码：

收录情况： SCI

摘要： Quality of dried mushroom slices (including colour, texture and chemical properties) is affected by drying temperature. In this study, NIR, MIR spectral data, colour features and texture features extracted from the images combined with machine learning models were used to identify the drying temperatures of facility cultivated Phlebopus portentosus mushrooms at 35, 45, 55, 65, and 75 degrees C with a view to providing a new method for traceability of agro-processed products. Under suitable pretreatments, the highest accuracy was 98.51% for the NIR-PLS-DA and NIR-SVM models, 91.04% for the MIR-PLS-DA model and 92.54% for the MIR-SVM model. Synchronous 2DCOS combined with ResNet models all had 100% accuracy. The LLDF and MLDF models can improve the performance of single-source dataset model. Feature extraction methods can reduce the risk of model overfitting. The 2DCOS images showed that distinguishing between Phlebopus portentosus mushrooms treated with different drying temperatures was mainly reflected in differences in the sugar, protein and ketone and alcohol feature bands, especially between 55 and 65 degrees C. These differences were associated with the Maillard reaction, but further research is needed to investigate and confirm the drying temperature that would allow the greatest retention of nutrients and the best flavour in dried mushroom slices. This study demonstrats the potential of NIR, MIR spectroscopy and image characterization for traceability monitoring of processed production of primary agricultural products. This initiative provides a green, high-throughput solution to enhance the quality of agricultural products and safeguard public health, and promotes the regulation of agro-processing production.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序