The invention discloses a method for quickly detecting
nutritional quality of root / stem crops on the basis of near-
infrared spectrum, and belongs to the field of
nutritional quality detection. The method includes steps of 1), screening a representative
test sample bank with high gradient of component content and according with modeling; 2), acquiring near-
infrared spectrum data of test samples by a near-
infrared spectrometer; 3), precisely analyzing component content of
nutritional quality of the test samples in the
test sample bank by referring to international or
national standard methods; 4),
processing the near-infrared spectrum data; 5), selecting a chemical metering process for building a
mathematical model and establishing a model by a program QUANT-2 in OPUS quantitative
analysis software; 6), evaluating the model by means of correction and internal
cross validation; and 7), collecting the near-infrared spectrum data of to-be-tested samples, preprocessing the spectrum data, and then predicting the nutritional quality of the samples by the aid of the established model. The method has the advantages of high analysis speed, no
pollution, no sample preprocessing, high detection precision, good
repeatability, and capability of analyzing various nutritional quality components of underground root / stem crops simultaneously.