期刊
JOURNAL OF FOOD PROCESSING AND PRESERVATION
卷 44, 期 5, 页码 -出版社
WILEY
DOI: 10.1111/jfpp.14421
关键词
-
资金
- Six Talent Peaks Project in Jiangsu Province [ZBZZ-019]
- Postgraduate Research & Practice Innovation Program of Jiangsu Province [KYCX17_1786]
- National Natural Science Funds Projects [31471413]
Moisture content was an important indicator to measure quality of green tea. In order to detect moisture content in green tea effectively and accurately, nondestructive detection for moisture content in green tea based on dielectric technology was proposed in this paper. Inductance-capacitance-resistance (LCR) measuring instrument and coaxial-line cylinder capacitor were used to collect dielectric data. The characteristic frequency points were extracted by successive projection algorithm (SPA) and variable iterative space shrinkage approach (VISSA). Support vector regression (SVR) was used to establish prediction models based on full frequency points and characteristic frequency points. The model results demonstrated that VISSA-SVR model based on dielectric loss factor epsilon '' performed best among all the prediction models, but the prediction accuracy was not enough, so the gray wolf optimization (GWO) algorithm was introduced to optimize the parameters (c and g) in SVR model. Furthermore, the best prediction performances for detecting moisture content in green tea was obtained, with the determination coefficient and root mean square errors (RMSEs) for prediction were 0.9695 and 0.0602, respectively. Therefore, dielectric technology combined with VISSA-GWO-SVR model is feasible for nondestructive determination of the moisture content in tea, which will provide a promising tool for the moisture content detection of other agricultural products. Practical applications Well understanding moisture content in tea is great importance. The practical application of this paper is to develop a novel method for moisture content detection in tea using dielectric technology. Compared with traditional methods, dielectric technology can be used to detect moisture content in tea nondestructively and accurately. Characteristic frequency selection algorithms are used to remove the redundant information in the data, and optimization algorithm is used to improve the performance of the model. Thus, dielectric technology combined with the optimal model is considered the most promising method for detecting the moisture content in green tea.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据