Post-harvest Storage Quality Inspection Of Kiwifruit Based On NIR And CT Techniques

The kiwifruit is one of the most important commercially cultivated fruit crops in the world.The post-harvest quality of kiwifruit fruit,which reaches edible conditions after ripening and softening during post-harvest storage,includes both chemical composition and microstructure,making it essential to construct non-destructive testing techniques for kiwifruit quality evaluation.In this paper,two types of kiwifruit,the yellow-fleshed(Xianwo2)and the green-fleshed(Cuiyu 38),were used to evaluate the main physicochemical parameters and tissue microstructure during post-ripening using near infrared spectroscopy(NIR)and X-ray CT imaging techniques,providing a new technical approach to the evaluation of kiwifruit for consumption.The main studies and results are as follows:NIR spectroscopy combined with chemometric methods was used to construct predictive models for storage period,single and mixed varieties of kiwifruit physicochemical indicators.The best preprocessing methods for NIR spectral data were obtained by comparative analysis as second-order differentiation and Norrise(25,5)smoothing.Partial least squares(PLS),reverse artificial neural network(BP-ANN)and support vector machine(SVM)were used to construct quantitative models for dry matter,brix and firmness of kiwifruit,respectively.The prediction regression coefficient(RP),root mean square error of prediction(RMSEP),and relative prediction deviation(RPD)were used to evaluate the model performance.When the RP is larger,the RMSEP is smaller,and the RPD is larger,the model performance is better.The RP,RMSEP,and RPD of the best models were as follows: 0.762,0.820,and 1.550(SVM)for dry matter,0.782,1.341,and 1.705(PLS)for brix,and 0.720,6.2501.501(PLS)for firmness,respectively.The dry matter model for Jade38 was 0.623,1.144,and 2.422(PLS);the sugar model was 0.658,1.261,and 2.561(PLS);and the firmness model was 0.663,2.072,and 3.707(PLS),respectively.The dry matter models for the mixed varieties were 0.740,0.964,and 2.005(PLS),respectively;the brix models were 0.7306,1.372,and 1.448(SVM),respectively;and the firmness models were0.619,6.156,and 1.198(SVM),respectively.The above results indicate that NIR can be used for postharvest index monitoring of kiwifruit and has the potential to develop a shared model for quantitative quality prediction of mixed varieties.(2)X-ray computed tomography combined with image processing was used to characterize the microstructural features of kiwifruit during storage.The average grayness of the body decreased from 680 to 560,inversely proportional to the storage time of the samples.The porosity increased from 0.12% to 2.1%,the pore volume from 20 mm3 to 300mm3 and the number of pores from 65 to 4000 in proportion to the storage time of the samples.The density fluctuated between 1.35 and 1.01 g/cm3 without significant changes.As storage time increased,kiwifruit softened after maturation.This eventually led to a decrease in the average ashiness of the body and an increase in the porosity,pore volume,and number of pores.The correlation of the three physicochemical indicators with microstructural characteristics was analyzed by correlation heat map.The correlation coefficients of average ash with firmness were positive up to 0.73 and negative up to 0.75 with brix.porosity was positively correlated with pore volume,number and porosity and positively correlated with brix up to 0.62 and negatively correlated with firmness up to 0.52.density was correlated with firmness up to 0.062 and correlation with physicochemical indexes existed.