[Determination of moistening endpoint of Phellodendri Chinensis Cortex based on multi-source data fusion]

This study aims to investigate the moistening mechanism of Phellodendri Chinensis Cortex via multiple characterization methods and characterize the moistening process and determine the endpoint based on multi-source data fusion. Low-field nuclear magnetic resonance(LF-NMR) was employed to investigate water migration and distribution patterns during the moistening process of Phellodendri Chinensis Cortex. The texture analysis was performed to monitor the hardness variations, and near-infrared spectroscopy(NIRS) was adopted to collect spectral data throughout the moistening process. The results showed that during the moistening process of Phellodendri Chinensis Cortex, water penetrated from the outer to the inner, and the content of semi-free water increased from 16.68% to 81.63% in the first 5 h. Meanwhile, the hardness decreased and gradually became stable after 5 h. The contents of berberine hydrochloride and phellodendrine hydrochloride changed mildly during the moistening process. The correlation analysis of LF-NMR parameters, hardness, water content, and chemical components demonstrated that total water content and semi-free water content exhibited correlations with hardness(P<0.05). Prediction models for water content and hardness were established based on near-infrared spectroscopy(NIRS). The models exhibited strong linear correlations(R~2>0.90), with root mean square errors of the calibration set and prediction set being closely aligned, indicating good robustness of the models. By integrating LF-NMR, texture analysis, and near-infrared spectral data, a prediction model for the moistening process and endpoint determination of Phellodendri Chinensis Cortex was established, providing a reference for non-destructive monitoring and endpoint determination in the moistening process of Chinese medicinal materials.