基于近红外光谱技术的藿香正气口服液质量透瓶快速检测方法研究

A study on the rapidly non-destructive detection method of Huoxiang Zhengqi oral liquid using near infrared spectroscopy

本文以厚朴酚、和厚朴酚、橙皮苷为关键质量指标,基于近红外光谱与高效液相色谱技术,建立藿香正气口服液近红外光谱与指标成分含量间的回归关系,从而实现基于近红外光谱技术的藿香正气口服液指标成分含量的透瓶快速测定。实验透瓶采集藿香正气口服液的近红外光谱,采用一阶导数、归一化进行光谱预处理,通过竞争性自适应重加权算法(CARS)筛选特征变量,采用偏最小二乘回归(PLSR)算法建立定量校正模型。然后,通过外部验证,对预测结果进行显著性检验。结果表明,透瓶扫谱建立的模型预测性能较好,验证集相关系数(R_(p))均高于0.99,预测均方根误差(RMSEP)均小于0.0084,验证集相对标准误差(RSEP)控制在2.83%以内;同时,透瓶模型和非透瓶模型的预测结果无显著差异。本文建立的透瓶检测方法具有模型性能好、预测精度高、快速、无损的优点,在藿香正气口服液质量控制中具有应用价值。

Based on near infrared spectroscopy and high performance liquid chromatography,this paper established the regression relationship between near infrared spectroscopy and index component content of Huoxiang Zhengqi oral liquid,so as to realize the rapid detection of index component content based on near infrared spectroscopy.Magnolol,honokiol and hesperidin were used as the quality indexes of Huoxiang Zhengqi oral liquid.After using the first derivative and normalization pretreatment method,characteristic variables were screened by CARS,and the correction model was finally established by partial least-squares regression(PLSR)method.The method accuracy was evaluated with the external validation,and the prediction results were tested for significance.The results indicated that when the near infrared spectrum was scanned through the bottle,the model’s correlation coefficients of prediction(R_(p))were higher than 0.99,the root mean square errors of the prediction models(RMSEP)were all less than 0.0084,and the relative standard errors of prediction set(RSEP)were all less than 2.83%.There was no significant difference in the predicted results between these two kinds of model.The models established in the non-destructive way have good performance and high prediction accuracy.The rapid and nondestructive way has application value in the quality control of Huoxiang Zhengqi oral liquid.