To reduce the influence of moisture content variation on the spectral detection of soluble solid content (SSC) and achieve rapid, non-destructive detection of SSC, as well as optimize the processing parameters of apricot products, improve detection efficiency and quality control, this study, based on spectral detection technology, combined with the multiple scatter correction (MSC) preprocessing method, concentration residual method for outlier removal, competitive adaptive reweighted sampling method for feature band selection and partial least squares (PLS) method, constructed SSC detection models for apricots at different moisture contents, and explored the impact of moisture content on the detection model effect of SSC in apricot. This study showed that as moisture content decreased, both the values of the valley depth and valley area in the first-derivative spectrum near 970 nm, as well as the peak height and peak area of the reflection peak, gradually diminished, reflecting the progressive weakening of moisture-related spectral characteristics. Furthermore, within the 450–1450 nm wavelength range, SSC prediction accuracy significantly improved with decreasing moisture content. Among the four moisture intervals, the optimal moisture range for apricot SSC detection is 30%-48%. The model exhibits an R p ² of 0.9690, an RMSEP of 0.5712°Brix, and an RPD of 5.7989, meeting the requirements for high-precision quantitative analysis. This study reveals variations in the hyperspectral SSC prediction performance across different moisture content ranges during the apricot drying process, providing a theoretical basis and technical reference for subsequent offline quality assessment and process monitoring of dried apricots.
Kang et al. (Fri,) studied this question.