Quality inspection of nectarine based on hyperspectral imaging technology

In this paper, the quality detection of nectarines based on hyperspectral imaging technology is proposed. The external indexes consist intact, cracked, rust, dysmorphic and dark damaged, while internal index composed soluble solid content (SSC). Firstly, 480 nectarine samples (160 intact 320 defective nectarines) with similar shape size are selected. Secondly, 5 spectral principal components 6 texture values acquired in range 420–1000 nm quality. Finally, methods Partial Least Squares (PLS), Support Vector Machine(LS-SVM) Extreme Learning Machine (ELM) used to establish discrimination models prediction models, respectively. As a result, accuracies 89.73%, 94.45% 88.62% obtained identification SSC predicted determination coefficients 0.8540, 0.8747, 0.8146, root mean squared errors 0.9849, 0.9101, 1.0732. results indicate great potential LS-SVM model predict discriminate inner outer nectarines.