Abstract

The variation in fertilization levels profoundly impacts the quality and astringency of persimmons (Diospyros kaki), making accurate assessment crucial for growers and consumers alike. Traditional methods for evaluating fertilization levels often involve destructive sampling, posing challenges in maintaining fruit integrity and sample throughput. In this study, we propose a nondestructive approach for assessing fertilization variation in persimmons using near-infrared (NIR) spectroscopy and nuclear magnetic resonance (NMR) relaxometry. We collected a diverse dataset of persimmon samples with varying fertilization levels and subjected them to NIR spectroscopy and NMR relaxometry analysis. NIR spectroscopy provided rapid and noninvasive measurements of chemical compositions, while NMR relaxometry offered insights into molecular dynamics and water content within the fruit tissues.

Multivariate data analysis techniques were employed to extract meaningful information from the spectroscopic data and develop predictive models for fertilization assessment. Partial least squares regression (PLSR) and other chemometric methods were utilized to correlate the spectral features with fertilization levels, enabling accurate and nondestructive quantification. Our results demonstrate the feasibility of using NIR spectroscopy and NMR relaxometry as complementary techniques for assessing fertilization variation in persimmons. The developed models exhibited high predictive accuracy, with potential applications in quality control, post-harvest management, and breeding programs. By providing a rapid and nondestructive means of evaluating fertilization levels in persimmons, this approach holds promise for enhancing fruit quality, optimizing production practices, and meeting consumer demands for consistently high-quality produce.