Apple Internal Quality Inspection Using Hyperspectral Image Technology

摘要

The internal parameters are important indexes for detecting the quality of the apples. This paper extracted spectral values of the apples from 400-1000nm with the hyperspectral image technology, carried out pre-treatment to original spectrums with MSC, performed regression analysis on spectral reflectivity of sugar content and firmness, and finally established prediction model of apple sugar content and firmness with BP (back propagation) artificial neural network. The results show that the correlation coefficient of the prediction model for sugar content is 0.9861, the average error is 0.118° Brix; the correlation coefficient of the prediction model for firmness is 0.9771, the average error is 0.054Kg/cm^2. Therefore, it is feasible to detect the internal quality parameter of apples using hyperspectral technology.