论文部分内容阅读
本文采集了40个复配乳油农药的近红外光谱,用偏最小二乘回归建立并优化了辛硫磷、高效氯氰菊酯和灭多威三种活性成分的定量分析模型。最优模型结果如下:辛硫磷模型的R2、RMSEC和RMSECV分别为99.53%、0.20和0.26;高效氯氰菊酯模型的R2、RMSEC和RMSECV分别为99.83%、0.06和0.07;灭多威模型的R2、RMSEC和RMSECV分别为99.76%、0.03和0.04。对三种成分进行了外部检验,辛硫磷、高效氯氰菊酯和灭多威的外部预测误差均方根(RMSEP)分别为0.31、0.12和0.08。结果表明,近红外光谱法可以用于复配乳油农药的活性成分含量的定量分析和质量控制,预测精度可以满足实际生产需要。
In this paper, the 40 near-infrared spectra of pesticides mixed with emulsifiable concentrates were collected. The partial least squares regression was used to establish and optimize the quantitative analysis models of three active components, phoxim, alpha-cypermethrin and methomyl. The results of the optimized model were as follows: R2, RMSEC and RMSECV of phoxim model were 99.53%, 0.20 and 0.26, respectively; R2, RMSEC and RMSECV of the beta-cypermethrin model were 99.83%, 0.06 and 0.07, respectively; R2, RMSEC and RMSECV were 99.76%, 0.03 and 0.04 respectively. External validation of the three components was carried out. The RMSEP of phoxim, beta-cypermethrin and methomyl were 0.31, 0.12 and 0.08, respectively. The results show that NIR spectroscopy can be used for the quantitative analysis and quality control of active ingredients in emulsifiable concentrate pesticides. The prediction accuracy can meet the actual production needs.