Fluorescence hyperspectral image technique coupled with HSI method to predict solanine content of potatoes

Abstract

In order to ensure the edibility of potatoes, fluorescence hyperspectral images of potato samples were obtained to predict the solanine content in potatoes. For the best ROI (region of interest), the S‐component of saturation was extracted by the HSI colorimetric technology to characterize the bud eye of potatoes in three‐dimensional geometric space. The effective bud eye was located as the geometric center of ROI and the average spectral information was obtained. After pretreatment and selection of feature wavelengths, the predicting mode of SVR was established and was optimized by adjusting the penalty coefficient c and the core coefficient g of radial basis function (RBF). Finally, the determinant coefficient of the model was 0.9143 and the root mean square error was 0.0296, which could basically meet the application requirements. It was concluded that the method based on hyperspectral fluorescence image and HSI colorimetry could predict the solanine content in potatoes accurately through the optimized SVR model. Practical applications The solanine content in potatoes is an important indicator for judging the edibility of potatoes. The traditional manual observation method has a high misidentification rate and is prone to waste. The chemical detection processes are also tedious and time‐consuming. Compared with traditional methods and chemical methods, hyperspectral imaging technology can detect the solanine content in potatoes rapidly, accurately, and nondestructively. In this study, we generated the best prediction model by SVR, which was established a base on characteristic wavelengths through best ROI selection. After comparing, it was found that the bud eye region is the best ROI and the spectral data obtained here has the best predictive ability. All these results provided a theoretical basis for the portable detection of solanine content in potatoes.

Current Application of Advancing Spectroscopy Techniques in Food Analysis: Data Handling with Chemometric Approaches

“…Potato is another food sample that has recently started to be useful for monitoring its quality with HSI. Lu et al assessed the impact of storage times on the evolution of solanine content in potatoes by using HSI in the spectral region of 500-1000 nm and support vector regression (SVR) and then allowed estimating the edibility of the potatoes [60]. Besides solanine content, the color is another indicator, that is also considered as a parameter to judge the quality of potatoes.…”

Hyperspectral image-based measurement of total flavonoid content of leaf-use Ginkgo biloba L.

“…Partial least square regression (PLSR) is a high-efficiency regression model in spectral analysis that combines the functions of multiple linear regression (MLR) and principal component regression (PCR), which can extract features among variables and analyze correlations among variables (Zhang et al, 2021;Chen et al, 2021). Support vector regression (SVR) is an important branch of support vector machine, that can transform a nonlinear problem into a linear problem (Tan et al, 2022;Lu et al, 2019). By minimizing support vector spacing, SVR can find the relationship between spectral information and target values.…”

Identification of fumigated and dyedLycium barbarumby hyperspectral imaging technology

“…When extracting spectral data, the selection of the region of interest (ROI) (Lu, Sun, Yang, & Hang, 2019) directly affects the accuracy of the classification model. Therefore, to collect the spectral data of a single sample more accurately, the entire L. barbarum specimen was chosen as ROI, and the average spectral data was extracted, as shown in Figure 3 (Yao et al, 2021).…”