| Fruit is favored by consumers because of its rich nutrition,and with the improvement of the quality of people's life,their requirements for the quality of fruit are also constantly improving.At present,the fruit industry has developed into an important industry in China.The detection of fruit quality is an important means to improve the value of fruit.In recent years,visible/near infrared(Vis/NIR)spectroscopy technology has shown great potential in the detection of fruit quality,mainly used to detect the soluble solids content(SSC),acidity,firmness,p H value and so on.Most of the studies are based on fresh fruits,and there are few studies on the quality of fruits during storage,but the rotting and damage of fruits caused by storage brings great economic losses every year.In addition,the size difference of fruit will affect the spectral collection during online detection,resulting in inaccurate spectral information,which will reduce the effect of the model in predicting fruit quality during storage.Therefore,the online detection equipment of fruit quality was used and the juicy peach was taken as the research object in this paper.The research on the detection of SSC and firmness of mengyin peach during storage by Vis-NIR spectroscopy was carried out,and the influence of peach size on the online detection was analyzed and the correction was carried out,thus realizing the quality detection of peach under different storage conditions.The main research contents and conclusions are as follows:(1)The influences of two detection methods on the quality model of peach in storage period were studied,and the effect of the two detection methods were compared.Taking juicy peach as the research object and the samples were stored at room temperature,then the spectra of samples were collected by diffuse transmittance and diffuse reflectance,and the SSC and firmness were used as evaluation indexes.The spectral analysis of the sample showed that the spectra collected by diffuse transmittance change obviously.The K-S algorithm was used to divide the samples into prediction set and calibration set,and the partial least squares(PLS)was used to develop the SSC and firmness models.At the same time,common preprocessing methods and wavelength selection algorithms were used to simplify or optimize the models.The results showed that the best prediction correlation coefficient(R_p)were 0.886 and 0.820,respectively,and the root mean square error(RMSEP)of SSC under diffuse transmittance and diffuse reflectance were 0.727°Brix and 1.003°Brix,respectively.The effect of SSC model under diffuse transmittance was better.The firmness models treated by uninformative variable elimination(UVE)and successive projections algorithm(SPA)were the best.In SPA-PLS model established by diffuse transmittance spectra,15 spectral variables were selected,and the R_p and RMSEP of firmness were 0.798 and 0.976 N,respectively.In the UVE-PLS model established by diffuse reflectance spectra,113 spectral variables were selected,and the R_p and RMSEP of firmness obtained were 0.841 and 0.829 N.It could be seen that both methods can detect the quality of peach during storage,and diffuse transmittance was generally better than diffuse reflectance,but the size difference will affect the accuracy of diffuse transmittance.(2)The effect of peach size on the diffuse transmittance spectra was studied and the size correction was carried out.The peach were divided into three categories according to their diameters:70-75 mm,75-80 mm and 80-85 mm,and their spectra were measured under four kinds of integration time.By analyzing the spectral data,it is found that the spectra of the latter two samples will decay under a short integration time,and the spectral information was incomplete.With the increase of the integration time,the attenuation value became smaller and smaller.The linear regression was used to process the spectral data under four kinds of integration time,and the slope of 70-75 fruit was 0.127,that of 75-80 fruit was 4.871,and that of 80-85 fruit was 5.656.It can be seen that the spectral intensity of the latter two changes greatly with the increase of integration time,and gradually approaches the maximum energy value.PLS was used to develop the SSC models of peach.64 prediction models were obtained by combining the three fruit types under different integration time.The results showed that when the integration time was 80-100-120 ms,the PLS model obtained by the combination of the three fruit types had the best effect.After multiplicative scatter correction pretreatment,the R_p and RMSEP of the model were 0.857 and 0.733°Brix,which indicated that the model could predict the SSC of three kinds of peach.The results showed that by changing the integration time,the size correction of peach could be realized,and the influence of fruit size on the online quality detection of peach could be reduced.(3)According to the detection methods and integration time obtained in the first two chapters,the quality changes of peach under different storage conditions were studied.The fruits were stored at room temperature and cold storage for 7 days,respectively.The spectra of the samples were measured by diffuse transmittance at the optimal integration time,and the SSC and firmness of the peach were measured.During storage,the spectral intensity of the peach increased,the wave peak slightly shifted,the SSC increased and the firmness decreased at room temperature.In cold storage,the spectral intensity decreases,the SSC decreased and the firmness was basically unchanged.In order to distinguish the spectral differences of peach under different storage days,principal component analysis and partial least squares discriminant analysis were used to develop the discriminant model.The misjudgment rate of the model was low,and the accuracy rate was higher than 90%.At the same time,the prediction model of SSC and firmness was developed by using PLS combined with a variety of pretreatment methods,and the optimal prediction model was obtained by UVE and competitive adaptive reweighted sampling(CARS).The results showed that PLS model was the best after treated by CARS.At room temperature,the R_p of SSC and firmness models were 0.819 and 0.811,respectively,and RMSEP were 0.841°Brix and 0.912 N,respectively.In cold storage,the R_p of SSC and firmness models were 0.828 and 0.785,respectively,and RMSEP were 0.816°Brix and 1.188 N,respectively.Finally,the correlation analysis of each index during the storage period showed that the SSC and firmness had negative correlation,and each index had a high correlation with the storage days,indicating that the storage days had a greater impact on the internal quality of the peach.The experiment showed that the model established after parameter optimization could well predict the quality change of peach under different storage conditions. |
PDF Full Text Request