Simultaneous And Online Detection Of Fruit Defect And Internal Quality Based On Visible-near Infared Tranmittance Spectroscopy

Fruit defects and the internal quality simultaneously on-line detection technology has important practical significance which can find external and internal fruit defects in a timely manner. At the same time, on-line detection technology improved the internal quality of fruits and the selling price which realize the target of raising market competitiveness. In foreign trade, it is beneficial to reduce the losses in the fruit export and improve the level of fruit industry additional value. Compared with the traditional damaged detection, visible-near infared tranmittance spectroscopy had a variety of advantages of using samples without damaged, easy to operate and on-line detection. To a great extent, the detection speed and accuracy rate were enhanced than before. The paper taken ‘Ya Li’ pear black-heart disease,amygdalus persica surface defect and citrus peel puffing disease as research objects to explore the feasibility of on-line simultaneously detection of fruit defects and the internal quality. This study could provide scientific reference and powerful technical support which made meaningful assistance to break the technology monopoly, reduce the return rates of foreign trade, increase fruit growers’ earning.The analysis of both experimental methods and conclusions of this study were described as below:(1) Visible/near-infrared diffuse transmission technology was developed to detect ‘Ya Li’pear black-heart disease applying qualitative discriminant models of peak ratio discrimination method, PCA and PLS-DA. Wherein the PLS-DA discriminant model for ‘Ya Li’ pear black-heart disease achieved the correct rate of 100%. At the same time, the PLS model of healthy pears eliminating black-heart disease pears was established to get the optimal coefficient with Rp was 0.93 and the RMSEP was 0.45°Brix.(2) Visible/near-infrared diffuse transmission technology was developed to detect amygdalus persica surface defect applying qualitative discriminant models of PCA,LS-SVM and PLS-DA. Wherein the PLS-DA discriminant model for amygdalus persica surface defect achieved the correct rate of 100%. At the same time, the PLS model of healthy amygdalus persica eliminating surface defect was established to get the optimal coefficient with Rp was0.95 and the RMSEP was 0.71 °Brix.(3) Visible/near-infrared diffuse transmission technology was developed to detect citrus peel puffing disease applying qualitative discriminant models of peak PCA, LS-SVM,BP-ANN and PLS-DA. Wherein the LS-SVM and PLS-DA discriminant model for citrus peel puffing disease achieved the correct rate of 100%. At the same time, the PLS model of healthy citrus eliminating peel puffing disease were established to get the optimal coefficient with Rpwas 0.88 and the RMSEP was 0.77 °Brix.(4) Using visible/near infrared diffuse transmission technology combined with chemometrics methods conducted the simultaneously on-line detection of fruit defects and the internal quality. The results found that the detection efficiency was more than 96%. The reaerch result showed that it is feasibility to conduct on-line detection of fruit defects and the internal quality simultaneously.