基于能量色散X射线荧光光谱的鱼粉产地溯源方法研究

doi:10.3964/j.issn.1000-0593(2021)03-0745-05

Abstract
References
Related Citation (15)

Download: PDF (2061 KB)
Export: BibTeX | EndNote (RIS)

Abstract Fishmeal is a kind of high protein feed material which plays an important role in aquaculture. There is a great market demand for fishmeal in China, but the quality of fishmeal from different places is different. In order to ensure the quality and safety of fishmeal, it is very important to establish a traceability system of fishmeal origin. The energy dispersion X-ray fluorescence spectrum is able to detect the type and content of mineral elements in the sample depending on the energy of the element’s radiation X-ray fluorescent photons. The types and contents of mineral elements contained in fishmeal may vary depending on the origin of fishmeal, so this paper proposes for the first time to use the energy dispersion X-ray fluorescence spectroscopy (EDXRF) method to scan the fishmeal to obtain the element information of fishmeal elements. After preprocessing the original spectrum, whale optimization algorithm is used to improve the adaptive net analyte signal weight K_lacal hyperplane method can identify the spectrum vector of fishmeal samples, and then identify the origin of fishmeal samples. Firstly, 51 fishmeal samples from Liaoning and Zhejiang were pressed, and different filters were set in the detection program of EDXRF spectrometer, 51 groups (6 spectra in each group) of spectra were obtained. Then the spectrum is preprocessed, and the baseline is corrected based on the adaptive iterative reweighted penalty least squares algorithm (airPLS), so as to eliminate the impact of baseline drift and improve the accuracy. Wavelet transform is used to smooth the spectrum and remove the high-frequency noise of the spectrum curve. The 16-dimensional vector representing the element content of each fishmeal sample was obtained by calculating the peak area of six effective spectral regions. Finally, whale optimization algorithm is used to select the key parameters (neighbor number, principal component fraction, adjustment parameter) of the adaptive net analyte signal weight K_local hyperplane (ANWKH) method, and then the adaptive net analyte signal weight K_local hyperplane model is established by using the found optimal parameters. 70% of the fishmeal samples from each place of origin are selected as the training set, 30% as the test set to identify the fishmeal place of origin. Fishmeal samples are from Liaoning and Zhejiang provinces. The accuracy of the prediction model is 94.3% and 100% respectively. The total accuracy is 97.3%, which is higher than the accuracy of the adaptive net analyte signal weight K_local hyperplane classification. The results show that the method based on energy dispersive X-ray fluorescence spectrum can accurately realize the origin traceability of fishmeal, and the adaptive net analyte signal weight K_local hyperplane method improved by whale optimization algorithm can find the optimal parameters and establish a model with higher classification accuracy. This paper provides a reference for more detailed origin traceability of fishmeal at home and abroad in the future.

Key words：Energy Dispersion X-ray fluorescence spectrum; Spectral pretreatment; Whale optimization algorithm; Adaptive net analyte signal weight K_local hyperplane; Fishmeal origin traceability

Received: 2020-03-30 Accepted: 2020-07-14

ZTFLH:

O657.34

Cite this article:

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2021)03-0745-05 OR https://www.gpxygpfx.com/EN/Y2021/V41/I03/745