Detection of different adulteration levels in red pepper using hyperspectral imaging combined with machine learning algorithms
Year: 2026
Type: article
Abstract: In this study, hyperspectral imaging was employed as an accurate and non-destructive method for detecting different adulteration levels in red pepper powder, a spice that is frequently targeted for economically motivated adulteration. Red pepper samples were deliberately mixed with chickpea flour, wheat flour, and sea foam powder at five adulteration levels (0, 5, 15, 30, and 50). Data were acquired using a line-scan hyperspectral imaging system in the visible to near-infrared range, and key wavelengths were identified through spectral analysis. Statistical features were extracted to be classified using artificial neural networks, random forest, and k-nearest neighbors methods. The accuracy of all algorithms in with feature selection strategy was substantially higher than without feature selection. Among the models, artificial neural networks achieved the highest accuracy, with values of 98.8, 100, and 97.8% for chickpea flour, sea foam powder, and wheat flour adulterants, respectively. These findings highlight the high potential of combining hyperspectral imaging with machine learning as an efficient and reliable approach for detecting adulterations in powdered food products, thereby contributing to improved food safety, quality control, and adulteration prevention.
Source: Sensing and Bio-Sensing Research
Institutions: Bu-Ali Sina University, Ilam University
Language: English
Subfield: Analytical Chemistry
Field: Chemistry
Domain: Physical Sciences
SDG: Zero hunger
Open Access status: gold
Estimated APC paid: $2,850
Funder: Bu Ali Sina University
OpenAlex ID: W7160084792
