Accurately determining citrus origin is essential for establishing and maintaining regional brands with distinctive qualities while safeguarding the rights and interests of both farmers and consumers. In this study, 2693 navel orange samples were collected from 13 major producing regions in China to establish a comprehensive near-infrared spectroscopy (NIRS) dataset. To address the challenge of citrus origin authentication, this study proposes a novel six-layer one-dimensional convolutional neural network (1D-CNN). The classification accuracy of this model reaches 96.16%. Compared with the support vector machine (SVM), partial least squares discriminant analysis (PLS-DA), and three-layer 1D-CNNs with kernel sizes of 3 and 16, the accuracy of the proposed six-layer model is improved by 9.65%, 3.21%, 3.84%, and 1.98%, respectively. Furthermore, the dataset is augmented using a Wasserstein Generative Adversarial Network (WGAN) and Noise Addition. The results indicate that data augmentation can effectively improve the accuracy of various algorithm models. Among them, the 1D-CNN proposed in this study achieves the best performance on the Noise Addition-augmented dataset, with its accuracy, precision, recall, and F1-score reaching 98.39%, 0.9843, 0.9839, and 0.9840, respectively. Compared with the other four comparative models, the accuracy of this model is increased by 1.48%, 1.36%, 1.48%, and 2.85%, respectively. Finally, a visual analysis of the 1D-CNN’s feature-extraction process was conducted. The results demonstrate that the 1D-CNN can effectively extract discriminative NIR spectral features to accurately distinguish citrus from different origins and that data augmentation markedly improves model performance by increasing data diversity. This work provides a robust tool for citrus origin tracing and offers a new perspective for the origin authentication of other agricultural products.

This study investigated the influence of harvest timing on the quality of navel oranges by the use of hand-held near-infrared spectroscopy (NIRS) device. ‘Newhall’ navel oranges were harvested in seven batches from 200 to 249 days after full bloom (DAFB). The samples were stored under cold storage (5 °C) and at room temperature (25 °C), respectively. Harvest-stage SSC and weight loss during storage were analyzed, and SSC-related trends in stored fruit were further evaluated at the 49-day endpoint. The results indicated that cold storage helped reduce weight loss in navel oranges. SSC exhibited a trend of increasing first and then decreasing with harvest date, reaching a peak at DAFB of 228. The NIRS model corrected by external parameter orthogonalization (EPO) achieved the best prediction performance for SSC, with R²p = 0.60, RMSEP = 1.38%, and RPD = 1.58 under room temperature, and R²p = 0.49, RMSEP = 1.46%, and RPD = 1.40 under cold storage. The EPO method reduced temperature-related interference and improved the robustness of SSC prediction under different storage conditions. Separately, based on measured postharvest weight loss and SSC changes, navel oranges harvested at 214–228 DAFB showed relatively better storage behavior under 5 °C conditions, with an estimated storage duration of approximately 62 days. Early-harvested fruit exhibited stronger storability but lower flavor quality, whereas late-harvested fruit had better eating quality but shorter storage life. It should be noted that these results correspond to different optimization objectives, including storability-oriented and quality-balanced criteria. This study provides preliminary evidence for the potential application of handheld NIRS in navel orange quality monitoring during storage, and offers an exploratory framework rather than a decision-making system for investigating the relationship between harvest timing and storage conditions. However, its predictive performance is moderate and further validation is required before practical application in production or decision-making systems.