[Objective] The paper was to explore a method for detecting chlorpyrifos residues in red Fuji apple. [Method] The original spectral data of apple samples sprayed with different volume fractions of chlorpyrifos were co...[Objective] The paper was to explore a method for detecting chlorpyrifos residues in red Fuji apple. [Method] The original spectral data of apple samples sprayed with different volume fractions of chlorpyrifos were collected using near infrared spectrometer at the band of 4 000-10 000 cm^（-1）. The original spectra were pre-treated by a variety of methods, and partial least squares（PLS） model was established for predictive analysis. [Result] Near infrared spectrum showed sensitivity to apple samples sprayed with different volume fractions of chlorpyrifos, but had low precision on pesticide-free samples. Data of blank control group were further eliminated for modeling prediction. The results showed that the results were the best when pre-treated with first derivative（FD）： R=0.987 9; the square error of prediction（SEP） was 0.173 6; the root-mean-square error of cross-validation（RMSECV） was 0.120 5; and the precision was 0.923 4. [Conclusion] Near infrared spectrum can better predict chlorpyrifos residue, providing a new method for detecting chlorpyrifos residues in Akesu red Fuji apple.展开更多
基金Supported by Emergency Management Project of National Natural Science Foundation of China(61640413)Open Project of South Xinjiang Agricultural Information Research Center of Agricultural Information Institute,CAAS(ZX2015005)Key Laboratory Project of Crop Water Use and Regulation,Ministry of Agriculture(FIRI2018-05-03)
文摘[Objective] The paper was to explore a method for detecting chlorpyrifos residues in red Fuji apple. [Method] The original spectral data of apple samples sprayed with different volume fractions of chlorpyrifos were collected using near infrared spectrometer at the band of 4 000-10 000 cm^（-1）. The original spectra were pre-treated by a variety of methods, and partial least squares（PLS） model was established for predictive analysis. [Result] Near infrared spectrum showed sensitivity to apple samples sprayed with different volume fractions of chlorpyrifos, but had low precision on pesticide-free samples. Data of blank control group were further eliminated for modeling prediction. The results showed that the results were the best when pre-treated with first derivative（FD）： R=0.987 9; the square error of prediction（SEP） was 0.173 6; the root-mean-square error of cross-validation（RMSECV） was 0.120 5; and the precision was 0.923 4. [Conclusion] Near infrared spectrum can better predict chlorpyrifos residue, providing a new method for detecting chlorpyrifos residues in Akesu red Fuji apple.
