| Sweet corn is one of the important new food crops.The acceptance grading and pricing of sweet corn are generally evaluated by sensory perception and experience.Traditional variety identification needs a lot of time,manpower and financial resources,and it is difficult to apply directly to the mass production process,so it is urgent to establish a scientific,quantitative and fast quality evaluation method.In this study,sensory evaluation of fresh food grains of 284 sweet corn inbred lines was conducted firstly.Then,chemical values of reducing sugar,sucrose and starch contents of 145 selected sweet corn inbred lines were determined.Finally,near-infrared spectroscopy was used to build a model of taste,reducing sugar,sucrose and starch.1,284 sweet corn inbred lines of fresh food grain seed coat thickness,according to taste,sweetness and tenderness of skin thickness,sweetness and tenderness of the mean value of 5.13,5.09 and 5.19 respectively,the variation coefficient were22.40%,22.83% and 21.88%,respectively,the seed coat thickness,sweetness and tenderness is divided into three levels,seed coat thickness 3 levels are 43 156 sweet corn,sweet corn,sweet corn in May;There were 38 sweetcorn,164 sweetcorn and 82 sweetcorn in the three grades.There were 29 sweet corn,165 sweet corn and 90 sweet corn in 3 grades of tenderness value.SPSS was used for variance analysis of 284 observed samples.The results showed that,at the significance level of 0.05,seed coat thickness,sweetness,tenderness and merchantability showed significant differences.With Num 10.0 as the threshold,the 284 sweet corn samples could be grouped into 5categories.The classification accuracy of three indexes including seed coat thickness,sweetness and tenderness combined with k NN model was 100% in the training set and 100%,98.25% and 100% in the test set.2.According to the rating of fresh food,145 sweet corn inbred lines were screened for the determination of glucose,sucrose and starch content.It was found that the average contents of reducing sugar,sucrose and starch were 0.91%,5.02%and 5.43%,respectively,with significant differences among varieties.The correlation coefficient between reducing sugar and sucrose,reducing sugar and starch was 0.16and-0.29,respectively,and the correlation coefficient between sucrose and starch was-0.43.The variation coefficients of reducing sugar,sucrose and starch content were63.13%,24.88% and 24.15%,respectively.SPSS was used for variance analysis of reducing sugar,starch and sucrose in 145 observed samples.The results showed that,at the significance level of 0.05,there were significant differences in the three dimensions.Correlation analysis of the three dimensions showed that reducing sugar was positively correlated with sucrose,but not significantly at 0.05 or 0.01(double tails).There was a significant negative correlation between sucrose and starch at level0.01(double-tailed).With Num 10.0 as the threshold,145 data can be roughly grouped into 5 categories,while 284 data can be roughly clustered into 5 categories.3.Six samples with abnormal spectral data were eliminated,and 139 sweet corn inbred lines were scanned by infrared spectrum for model construction.The original spectral data of 139 sweet corn inbred lines were pretreated with Ram,MSC and SNV,indicating that the optimal results of reducing sugar,sucrose and starch could be obtained by SNV pretreatment.R2 was 0.95,0.99 and 0.96,respectively,showing strong correlation.The prediction formulas were Y=1.01x-0.028,Y=0.98x+0.054 and Y=0.95x+0.026,respectively. |
PDF Full Text Request