ASHS 2024 Conference

Sweet potato (Ipomoea batatas L.) is one of the seven major food crops grown worldwide. The optimal storage temperature for sweet potatoes is 12-15℃, and chilling injury occur when it falls below 10℃. The initial symptom of chilling injury in sweet potatoes is difficult to discern in visual judgment and requires checking through a cut-off sections. Therefore, this study intended to develop a non-destructive chilling injury prediction model using VIS/NIR spectroscopy sensors. USB4000 Fiber Optical Spectrometer was used for collecting spectrum data in the wavelength of the visible to near-infrared region (400 nm to 1100 nm), and the average value was used by measuring six areas divided into upper, middle, and lower parts to reduce errors. Firmness, total soluble solids, internal chromaticity, chilling index, and Malondialdehyde (MDA) were measured as reference data, and Partial Least Squares Regression(PLSR) method was used for prediction. To increase the accuracy of the model, the near-infrared spectrum data were preprocessed using the Savitzky–Golay 1st and 2nd derivatives (S–G), standard normal variate (SNV), and multiplicative scattering correction (MSC) methods. Excellent performance models for each preprocessing methods were selected, and as a result of verification through the prediction model, the 'Hogammi' variety showed excellent performance in the preprocessing of Savitzky-Golay 2st derivative in TSS and Malondialdehyde (MDA), and CI index showed the best performance in the processing of SNV. As indicators of chilling injury such as the CI index involve subjective opinions of experimenters, the findings of this study could be used as a fundamental step to develop more objective and accurate chilling injury prediction models by including various internal indicators.