Abstract Total volatile basic nitrogen (TVB-N) content is an important indicator for evaluation of duck-meat freshness. The accuracy of the existing near-infrared spectrometric model is affected by sample characteristics of different varieties. Hence, improvement in the model is needed. In order to improve the applicability of the model for TVB-N content in duck meat with different varieties, calibration maintenance methods were done, based on different principles including model updating (MU), direct standardization (DS), and slope/bias correction (SBC). The numbers of samples in the standardization sets were investigated for transfer effects. The results showed that the univariate SBC method exhibited a more significant transfer effect than multivariate DS transfer or complete recalibration of MU. The best results were obtained by the SBC method with 20 standard samples, for which the root mean square error of prediction and mean bias were 0.930 and –0.130. The calibration maintenance and transfer study done will likely provide reference for model maintenance between samples with different varieties and lay a foundation for industrial application.

中文翻译：

---

基于近红外光谱模型校准维护和传递的不同品种鸭肉总挥发性碱性氮含量

摘要 总挥发性碱性氮（TVB-N）含量是评价鸭肉新鲜度的重要指标。现有近红外光谱模型的准确性受不同品种样品特性的影响。因此，需要对模型进行改进。为了提高模型对不同品种鸭肉中TVB-N含量的适用性，基于模型更新（MU）、直接标准化（DS）、斜率/偏差校正等不同原则，进行了校准维护方法（ SBC）。研究了标准化集中的样本数量的转移效应。结果表明，单变量 SBC 方法比多变量 DS 传递或 MU 的完全重新校准表现出更显着的传递效果。最好的结果是通过 SBC 方法获得的 20 个标准样本，其预测的均方根误差和平均偏差分别为 0.930 和 –0.130。完成的校准维护和转移研究可能为不同品种样品之间的模型维护提供参考，为工业应用奠定基础。