近红外光谱技术在快速检测辣椒粉吸光度中的应用研究

Application Research on Rapid Detection of Chili Powder Absorbance by Near Infrared Spectroscopy Technology

为提高辣椒粉吸光度的检测效率,以新疆辣椒粉为建模材料,使用波通、蔡司和赛默飞3个品牌的近红外光谱仪对辣椒粉进行扫描建模,确定最佳方式为赛默飞近红外光谱仪在波数精度16 cm^(-1)条件下采集光谱,利用偏最小二乘法建立模型。经过一阶导数+SG平滑两步光谱预处理方式,并选择主因子数为15,对模型进行优化。校正均方根误差(Root Mean Square Error of Calibration,RMSEC)为0.218,校正相关系数(Correlation Coefficient of Calibration,Rc)为0.9934;日内精密度(n=3)最大为1.22%,日间精密度(n=3)最大为0.98%,93%的样品化学实验值与近红外模型预测值两者之差在0.6以内,结果表明该近红外模型稳定性好,精确度高,能实现辣椒粉吸光度的快速检测。

In order to improve the detection efficiency of absorbance of chili powder,Xinjiang chili powder was used as modeling material,and near infrared scanning model of chili powder was established by near infrared spectrometers of Perten,Carl Zeiss and Thermo Fisher.The near infrared ray model was best under the wave number accuracy of 16 cm^(-1),and the model was established by partial least square method.The model was optimized by the first derivative+SG smoothing two-step spectral preprocessing method and the principal factor number as 15.The root mean square error of calibration(RMSEC)was 0.218.The correlation coefficient of calibration(Rc)was 0.9934.The maximum intraday precision(n=3)was 1.22%and maximum daytime precision(n=3)was 0.98%.The difference between the chemical experimental values and the predicted values was within 0.6 of the 93%samples.It indicated that the near infrared model has good stability and high accuracy,and it can achieve rapid detection of absorbance of chili powder.