Research on Nondestructive Detection of Hawthorn Quality Based on Hyperspectral Imaging Technology

Abstract

Hawthorn, a unique medicinal and edible economic crop in China, is prone to damage and pest infestations during harvesting and transportation, which severely degrade its quality and utilization value. Traditional manual and instrumental detection methods are characterized by low efficiency, strong subjectivity, and destructiveness, failing to meet the demands of large-scale quality inspection. This study focuses on hawthorn as the research object and conducts non-destructive detection of its damage and pests using hyperspectral imaging technology. By reviewing the domestic and foreign research progress in non-destructive detection of fruit damage, the application potential of hyperspectral imaging technology in the micro-damage and microcosmic detection of agricultural products is clarified. A technical route of “sample preparation—hyperspectral image acquisition—spectral preprocessing—model construction and validation” is designed. Spectral preprocessing is implemented via methods such as Standard Normal Variate (SNV) and Savitzky-Golay (SG) smoothing, and a detection system is established by integrating models including Partial Least Squares (PLS), Principal Component Regression (PCR), and Least Squares Support Vector Machine (LS-SVM). This study aims to address issues such as high sample misjudgment rate and redundant spectral information in hawthorn damage, with the expectation of achieving rapid, non-destructive, and accurate recognition. The research findings can provide technical support for quality grading and non-destructive detection in the hawthorn industry, as well as theoretical and practical references for the application expansion of non-destructive detection technology in agricultural products.