Construction Of Near Infrared Model Of Total Soluble Sugar In Fresh Grain Of Super Sweet Corn And Its Contrastive Analysis

In this experiment,170 sweet corn hybrids at home and abroad were screened artificially.The fresh grains of this population were scanned by spectroscopy,and the chemical indexes of total soluble sugar were detected.The results of spectroscopy scanning were correlated with the chemical values obtained by analysis software.The results were applied in practical production to provide rapid and accurate identification indicators for breeding work.The study found that:1.Establishment of Near Infrared Model of Total Soluble Sugar in Fresh Seeds of Super Sweet CornThe correlation coefficient of the NIR model of soluble total sugar in fresh grains was 0.8204 after the first derivative(1der)plus 7-point window smoothing(SW),and the correlation degree of the model was high,showing a strong correlation.The results show that the first derivative can be used to construct the model of soluble total sugar in near infrared spectroscopy after processing spectral information.When the whole spectrum was divided into 30 intervals and four intervals[1,5,9,27],the soluble total sugar model was the best.RMSEV = 5.027,RC = 0.825.The model was used to predict the soluble total sugar content in sweet corn fresh grains.The predicted results were RMSEP value of 4.504 and Rp value of 0.862.The results show that the model constructed by 1der plus SW 7-point smoothing can satisfy the needs of routine detection when the whole spectrum is divided into 30 intervals and four intervals [1,5,9,27] are combined.2.Establishment of Mid～Infrared Model of Total Soluble Sugar in Fresh Seeds of Super Sweet CornAfter standard normal transformation(SNV)and subtraction trend(TD),the mid-infrared spectroscopy model of total soluble sugars in fresh grains was highly correlated,reaching 0.8184,showing a strong correlation.The results showed that SNV plus TD could be used to construct the soluble total sugar model of mid-infraredspectroscopy.The total spectrum is divided into different intervals and the number of joint sub-intervals,which have an effect on the mathematical modeling.The total spectrum is divided into 30 intervals.When two intervals [8,27] are combined,the total soluble sugar model is the best.RMSEV = 5.369,RC = 0.818.Finally,the selected model is used to compare the total soluble sugar content in fresh sweet corn grain,and the selection prediction is more effective.Accurate model.Finally,RMSEP=4.646 and Rp=0.867 were selected as the prediction results.From the prediction results,we can see that the correlation coefficient of the spectral prediction is 0.8721,which can be applied to routine sample detection,but the established spectrum needs to be supplemented and modified to improve the accuracy of the model.The results show that near infrared spectroscopy combined with first derivative processing,mid-infrared spectroscopy combined with standard normal transformation and de-trend processing can effectively overcome some of the effects caused by external interference,and get a high degree of correlation modeling results.The experiment proves that near infrared spectroscopy can provide a rapid and non-destructive quantitative analysis method for the determination of total soluble sugar in sweet corn fresh grains.