In this work, the near infrared spectroscopy (NIR) technology was applied to nondestructively evaluate the freshness of pork. The total volatile basic-nitrogen (TVB-N) and pH value of pork were detected as freshness evaluation indicators. A multi-index statistical information fusion (MSIF) modeling method based on variable selection was proposed to evaluate pork freshness. In the experiment, the proposed MSIF was compared with other state-of-the-art variable selection methods. Results showed that the proposed method achieved the best generalization performance and stability. The prediction correlation coefficient (Rval) and root mean square error (RMSEP) of MSIF were: Rval = 0.8618 and RMSEP = 3.910 for TVB-N content, Rval = 0.9379 and RMSEP = 0.1046 for pH value. The research demonstrated that NIR combined with MSIF has the potential for rapid and nondestructive determination of pork freshness, and so hopefully to provide a promising tool for monitoring meat quality and enriching the extracted information from food industry.

在这项工作中，将近红外光谱（NIR）技术用于无损评估猪肉的新鲜度。检测猪肉的总挥发性碱性氮（TVB-N）和pH值作为新鲜度评估指标。提出了一种基于变量选择的多指标统计信息融合建模方法，以评价猪肉的新鲜度。在实验中，将拟议的MSIF与其他​​最新的变量选择方法进行了比较。结果表明，该方法具有最佳的泛化性能和稳定性。MSIF的预测相关系数（Rval）和均方根误差（RMSEP）为：TVB-N含量的Rval = 0.8618和RMSEP = 3.910，pH值的Rval = 0.9379和RMSEP = 0.1046。