Work

Combining VNIR and SWIR hyperspectral imaging with data fusion for rapid and nondestructive detection of soluble solids content in cherries

Year: 2026

Type: article

Abstract: Soluble solids content (SSC) serves as a crucial indicator for evaluating the quality of cherries. This study aimed to achieve an accurate and non-destructive prediction of SSC in cherries. Hyperspectral images of Cherry were acquired in both the visible near-infrared (VNIR) and short-wave infrared (SWIR) spectral ranges. To address the issue of high-dimensional spectral features, multiple feature extraction algorithms were introduced to reduce data redundancy and improve modeling efficiency and prediction accuracy. Based on this, three data fusion strategies, including low-level, mid-level, and high-level fusion, were developed to integrate VNIR and SWIR information. Partial least squares regression (PLSR) was used as the modeling algorithm to establish the SSC prediction model. The results demonstrated that all multi-level fusion strategies effectively enhanced model performance, with high-level fusion showing the best results across various evaluation metrics achieving an R² p of 0.8385, RMSEP of 0.5246 °Brix, confirming the feasibility and advantages of synergistic modeling using multi-spectral information. In addition, The PLSR model based on HSI data was used to visualize the spatial distribution of SSC in cherries.

Source: Journal of Food Composition and Analysis

DOI: 10.1016/j.jfca.2026.109376

Authors: Huaming Liu, Ruihao Wu, Xiuyou Wang, Xuehui Bi, Peng Li

Institution: Fuyang Normal University

Language: English

Topic: Spectroscopy and Chemometric Analyses

Subfield: Analytical Chemistry

Field: Chemistry

Domain: Physical Sciences

SDG: Zero hunger

Open Access status: hybrid

Estimated APC paid: $3,450