| Some kinds of fruit are damaged easily by mechanical damage in the process of picking,transporting and sorting.Deficiency and soluble solid content(SSC)are important indexes for evaluating the quality of fruits.Fruits deficiency can easily lead to bulk return phenomenon in foreign trade.It will cause huge economic losses.The surface scratches and bruises were cause easily when sorting fruits which easy to damage and with high water content online.The feasibility was discussed for nondestructive detecting with net for SSC and simultaneous detecting deficiency and SSC of fruits by online visible-near infrared(visible-NIR)transmittance spectroscopy.The paper taken ‘Ya Li' pear,dangshan pear and hosui pear as research objects to recorded visible-near infrared(visible-NIR)transmittance spectroscopy.The samples were conveyed at the speed of five samples per second.Ten tungsten halogen lamps were installed in a sorting line.The power of each lamp was 100 watt.The spectra of sample were analyzed for investigation of the characteristics of spectral.The robust model was developed in the finally.The analysis of both experimental methods and conclusions of this pear were expressed as below:(1)The ‘Ya Li' pear was used as research objects to discuss for detecting blackheart,pest and SSC of intact ‘Ya Li' pear simultaneously by online visible-near infrared(visible-NIR)transmittance spectroscopy.Discrimination models of principal components analysis(PCA),least squares support vector machine(LSSVM)and discrimination partial least square(DPLS)were developed for discrimination of blackheart,pest and normal fruits.DPLS was superior to PCA with relative higher classification rate of 99.2%,and suitable for online sorting compare with LSSVM.PLS model was employed with normal pears.Therefore a new strategy was proposed for online sorting for intact pears simultaneously by online visible-NIR transmittance spectroscopy.(2)The paper taken dangshan pear as research objects to investigated for detecting injured pear and SSC of intact dangshan pear simultaneously by online visible-near infrared(visible-NIR)transmittance spectroscopy.Three methods of principal components analysis(PCA),correlation analysis and discrimination partial least square(DPLS)were developed for discrimination of injured pear and sound pear.DPLS was superior to PCA with relative higher classification rate of 100%,and suitable for online sorting compare with LSSVM.Then the PLS regression model was developed for predicting SSC values.The accuracy of sorting grade was 96.7% according to the SSC values.(3)The feasibility was discussed for nondestructive detecting with net for SSC of intact hosui pear by online visible-near infrared(visible-NIR)transmittance spectroscopy.The influence of pears with net to SSC determination was also explored by partial least square(PLS)regression models,and PLS model was employed with normal pears and pears with net.The robust model was developed by the spectra which eliminate the influence of net by binomial fitting.Therefore a new strategy was proposed for detection of SSC for intact pears by online visible-NIR transmittance spectroscopy.The new samples,which were not used in the calibration,were used to access the abilities of predicting SSC of intact pears.The accuracy of sorting grade was 94.4% according to the SSC values and the SEP was 0.328% for predicting SSC.The results showed that nondestructive detecting with net for SSC of intact hosui pear were feasible by visible-NIR transmittance spectroscopy.The research can provide technical support and reference basis for grading of mass fruits quality.The research can provide technical support and reference basis for grading of mass fruits quality. |
PDF Full Text Request