| Fruits and vegetables as an important part of residents'diet are rich in nutrients.However,postharvest fruits and vegetables are highly prone to water loss,aging and deterioration,which would ultimately result in the loss of freshness.Freshness detection can effectively reduce the freshness loss of fruits and vegetables and ensure their postharvest qualities.Nevertheless,there are many kinds of fruits and vegetables with various quality indicators.Traditional freshness evaluation methods based on some of these indicators is time-consuming,and cannot accurately reflect the freshness and quality of fruits and vegetables.With the development of agricultural economy and the advancement of science and technology,intelligent detection method is becoming the trend of the food industry.It is of great significance to carry out sensor-based intelligent detection of freshness to ensure food production and promote the development of food industry.Therefore,the correlations between flavor characteristics/water statue and storage quality of two typical vegetables(cucumbers and cherry tomatoes)during storage were researched in this study.On this basis,the models of fruits and vegetables freshness detection based on electronic nose and low field NMR were constructed by means of stoichiometry,and the applicability of the models were evaluated.Furthermore,a"fruits and vegetables freshness intelligent detection"system was developed based on R language,achieving the intelligent identification of freshness and quality of fruits and vegetables.The specific results are as follows:The changes of flavor characteristics and water status of cucumbers and cherry tomatoes during storage were analyzed,and the correlation between flavor characteristics/water status of cucumbers and cherry tomatoes and their storage qualities was studied.The results of sensory evaluation showed that:1)Cucumbers can be divided into four different freshness grades:fresh,medium fresh,acceptable and spoiled during storage.Cherry tomatoes can be divided into three different freshness grades:fresh,acceptable and spoiled.2)The hardness,soluble solids,pH and color difference of cucumbers and cherry tomatoes changed significantly with storage time.Cluster analysis based on the above quality indexes showed that the cucumbers and cherry tomatoes can be divided into the same classification group as the result of sensory evaluation.3)During storage,the flavor characteristics and water status of cucumbers and cherry tomatoes varied with the change of storage quality.Correlation analysis showed that there was a significant or extremely significant correlation between sensory scores,quality indicators,flavor characteristics and water status during storage of cucumbers and cherry tomatoes.Among them,the storage quality index and flavor characteristics of cucumbers and cherry tomatoes have strong correlation,followed by the correlation between the storage quality index of cucumbers and water status.Therefore,flavor characteristics and water status can be used to evaluate the quality changes of fruits and vegetables during storage.According to the correlation between flavor characteristics and quality of cherry tomatoes,partial least squares(PLS)and support vector machine(SVM)methods were used to construct the freshness classification recognition model and quality prediction model,respectively,based on the electronic nose data.The results showed that:1)Both PLS and SVM can effectively discriminate the freshness grades(fresh,acceptable and spoiled)of cherry tomatoes.The accuracy of the two methods for the freshness detection of cherry tomato is 100%.2)In the prediction of quality index,the test set R_p~2 of the PLS model for firmness,pH,SSC and total color difference of cherry tomatoes was 0.9079,0.9323,0.9249,0.9826,respectively,RMSEP was 0.5399 N,0.0247,0.1613%,0.3687,respectively,and RPD was 3.3505,3.9070,3.7108,7.7213,respectively.Therefore,the electronic nose analysis combined with the PLS model can effectively predict the freshness and quality of cherry tomatoes during storage.Based on the correlation between storage quality and flavor characteristics as well as the water status of cucumbers,PLS and SVM were used to establish the freshness classification model and quality prediction model with electronic nose and LF-NMR data,respectively,and the prediction performance of the models were compared.Results showed that:1)compared with the electronic nose,the accuracy of model established by LF-NMR combined with SVM was 96.67%,in which the rate of discrimination accuracy of freshness,medium freshness,acceptability and spoilage were 100.00%,90.00%,100.00%and 100.00%respectively,which improved the accuracy of freshness identification of cucumber;2)the prediction models for cucumber quality indexes,including firmness,pH,SSC and total color difference,gave better performance with R_p~2of 0.8897,0.8758,0.9245,0.8955,RMSEP of 0.8044 N,0.2473,0.1523%,1.0585,and RPD of 2.7093,2.7947,3.3189,2.9872,respectively.Therefore,models built by LF-NMR analysis combined with SVM can predict the freshness and quality of cucumbers very well.In order to evaluate the applicability of the models,the above models of freshness and related qualities were used to detect the postharvest qualities of cherry tomatoes subjected to high pressure argon and cucumbers treated with ultrasound respectively.The results showed that:1)high pressure argon treatment,especially at 0.8 MPa could effectively inhibit the decrease of firmness and increase of total color difference,reduce the change rate of pH and soluble solids,and maintain the freshness of cherry tomatoes better.2)Ultrasound treatment could also inhibit the decrease of firmness and increase of total color difference,reduce the change rate of pH and soluble solids,maintain the freshness of cucumbers better.Effects of high pressure argon and ultrasound treatment on the preservation of cherry tomatoes and cucumbers can be identified by the above models accurately.The relative errors between the predicted values and measured values of firmness,pH and soluble solids of cherry tomatoes and cucumber were less,and the identification rates for freshness of cherry tomatoes and cucumbers were more than 85%.It can be seen that the above models have a wide range of applicability in the detection of freshness and storage quality of fruits and vegetables under different treatment conditions.On the basis of established freshness and quality detection models,an intelligent recognition system for freshness of fruits and vegetables was developed based on R language.This system can be used to detect the freshness and quality of cherry tomatoes and cucumbers intelligently by introducing electronic nose response value or LF-NMR signal.In addition,this system also has the functions of continuous supplement of the database and optimization of the models to realize the intelligent detection of a variety of fruits and vegetable as well as various quality indicators.The friendly interface,simple operation,powerful function and intuitive display of the system initially realizes the fast and non-destructive intelligent detection of the freshness of fruits and vegetables. |
PDF Full Text Request