Snapshot-Based Multispectral Imaging for Heat Stress Detection in Southern-Type Garlic

This study aims to develop a model for detecting heat stress in southern-type garlic using multispectral snapshot camera. Raw images were obtained from cloves during the bulb enlargement period, capturing visible (Vis) and near-infrared (NIR) regions. Image preprocessing was applied obtain 38-wavelength spectrum by combining 16-wavelength image Vis region 22-wavelength NIR region. These spectral data then utilized models, including PLS-DA, LS-SVM, DNN, recurrence plots-based CNN (RP-CNN). On average, LS-SVM demonstrated best performance period. is attributed nonlinear nature of differences between groups caused abiotic garlic. The particularly effective at such relationships. Among images, yielded performance, followed RP-CNN, PLS-DA. Therefore, this confirms potential snapshot-based imaging measuring changes crops induced high-temperature stress.