2026 年 64 巻 1 号 p. 1-12
This study aimed to develop a non-destructive method for measuring potassium content in leafy vegetables to support quality assurance of low-potassium vegetables for patients with chronic kidney disease. The test subjects were komatsuna, bok choy, wasabi leaf, and celtuce. Reflectance spectra in the visible (400―700 nm) and near-infrared (900―1700 nm) regions were obtained using a portable spectrometer and analyzed using second derivative preprocessing. Actual potassium content was measured using an inductively coupled plasma emission spectrometer. Calibration models were developed using partial least squares regression to predict potassium content. The correlation coefficients between predicted and measured potassium values were 0.843, 0.960, 0.927, and 0.915 for komatsuna, bok choy, wasabi leaf, and celtuce, respectively, indicating strong predictive performance. The residual predictive deviation (RPD) values were 1.57, 3.43, 2.47, and 2.26, respectively. These results confirm the feasibility of non-destructive potassium measurement using visible and near-infrared spectroscopy. This approach enables analysis of entire sample batches, enhancing quality assurance while reducing sample loss and cost. It holds promise for promoting the use of low-potassium vegetables and improving production efficiency.
