Agricultural Product Processing and Storage Lab, School of Food and Biological Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China.
College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310058, China.
Food Chem. 2023 Jun 15;411:135431. doi: 10.1016/j.foodchem.2023.135431. Epub 2023 Jan 7.
Analogous density foreign matter (ADFM) embedded in soy protein meat semi-finished (SFSPM) is hidden by SFSPM and has similar acoustic impedance features to SFSPM, which makes non-destructive testing techniques such as computer vision (CV), reflectance spectroscopy and ultrasound imaging inappropriate for ADFM, which not only seriously affects the quality of soy protein meat (SPM) products but also increases the safety risk to consumers. In this study, to locate and separate ADFM by using transmission hyperspectral imaging (T-HSI) technique which is sensitive to chemical composition and highlight internal contours. The optimal discrimination model SVM + PCA + MSC + SPA was constructed using transmission spectral information with an accuracy of 95.00 %. The visualization results based on the optimal model showed clearer localization results than CV and ultrasound imaging. The study demonstrated that the advantages of T-HSI technology in detecting and locating ADFM inside SFSPM, which provides a basis for improving the production quality and safety of SPM.
类似密度异物(ADFM)嵌入在大豆蛋白肉半成品(SFSPM)中,被 SFSPM 隐藏,且具有与 SFSPM 相似的声阻抗特性,这使得计算机视觉(CV)、反射光谱和超声成像等无损检测技术不适用于 ADFM,这不仅严重影响大豆蛋白肉(SPM)产品的质量,还增加了消费者的安全风险。在这项研究中,为了通过对化学成分敏感并突出内部轮廓的透射高光谱成像(T-HSI)技术来定位和分离 ADFM。使用透射光谱信息构建了最优判别模型 SVM+PCA+MSC+SPA,准确率为 95.00%。基于最优模型的可视化结果显示出比 CV 和超声成像更清晰的定位结果。该研究表明,T-HSI 技术在检测和定位 SFSPM 内 ADFM 方面具有优势,为提高 SPM 的生产质量和安全性提供了依据。