Convolutional Neural Networks of Whole Jujube Fruits Prediction Model Based on Multi-Spectral Imaging Method

Soluble sugar is an important index to determine the quality of jujube, and also factor influence taste jujube. The acquisition soluble content jujube mainly relies on manual chemical measurement which time-consuming labor-intensive. In this study, feasibility multispectral imaging combined with deep learning for rapid nondestructive testing fruit internal was analyzed. Support vector machine regression model, partial least squares convolutional neural networks (CNNs) model were established by method predict whole fruit, optimal selected three kinds sugar. study showed that sucrose prediction had best performance after CNNs training, correlation coefficient verification set 0.88, proved using fruits.