首页|Hyperspectral detection of salted sea cucumber adulteration using different spectral preprocessing techniques and SVM method
Hyperspectral detection of salted sea cucumber adulteration using different spectral preprocessing techniques and SVM method
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Elsevier
Salted (S) sea cucumbers are not only used as the ingredient in most sea cucumber products but also sold directly as the final product. Over-salted (OS) and sugar-treated (ST) sea cucumbers are usually used to adulterate salted sea cucumbers, which jeopardizes the quality and safety of salted sea cucumber products. In this study, S, OS and ST sea cucumbers were classified according to the hyperspectral data in the near-infrared range (957.3-1679.8 nm). Four preprocessing methods, including Savitzky-Golay (SG), multiplicative scatter correction (MSC), standard normal variate (SVN) and variable sorting for normalization (VSN) were established to filter the noise and scattering information in the original spectra. The approaches of regression coefficient (RC), successive projections algorithm (SPA), and two-dimensional correlation spectroscopy (2D-COS) were adopted to extract characteristic wavelengths. Subsequently, sea cucumber classification models were established on the basis of full and characteristic spectra using a support vector machine (SVM). Among the models established, the SNV2D-COS-SVM and VSN-2D-COS-SVM classification models showed optimal recognition performance. In summary, the rapid and non-destructive classification and identification of S, OS, and ST sea cucumbers were accurately realized with the near-infrared-based HSI technology, and the feasibility of using near-infrared hyperspectral technology to identify salted sea cucumber was verified.
Classification and recognitionHyperspectrumSea cucumberSupport vector machine
NSTL
Elsevier
