近红外光谱检测青贮饲料的营养成分

Study on the nutritional components of silage with Near Infrared Spectroscopy

选取154批次青贮样品为研究对象,利用化学分析法测定营养成分的含量,采用近红外光谱仪对其采集光谱数据,应用化学计量学软件中的偏最小二乘法(PLS)建立青贮饲料化学成分的定标模型,并探讨定标模型的可行性。定标模型的决定系数(R2)分别为0.913,0.854,0.992,0.808,0.884,0.870;标准偏差(SEC)分别为0.322,0.215,0.230,1.909,2.785,2.275。验证化学分析值与预测值之间的决定系数r2分别为:0.9544,0.779,0.9604,0.8031,0.8539,0.8227。结果表明,近红外光谱分析技术可用于青贮中粗蛋白、水分、淀粉、中性洗涤纤维、酸性洗涤纤维含量的定量检测分析。

The 154 silage samples were grounded,using the chemical method to analyz nutritional composition,using the near spectrometer to scan the spectrums.All the NIRS calibrations for the silage were developed with partial least squares(PLS)regression.R 2 and SEC were 0.913 and 0.322,0.854 and 0.215,0.992 and 0.230,0.808 and 1.909,0.884 and 2.785,0.870 and 2.275;r2,the identifying coefficient between chenical analysis results and predicted results was 0.9544,0.779,0.9604,0.8031,0.8539 and 0.8227.So,near-infrared prediction model was feasible to measure crude protein,moisture,starch,NDF and ADF in corn silage quickly.