Rapid Identification of Apple Maturity Based on Multispectral Sensor Combined with Spectral Shape Features

The rapid and convenient detection of maturity is great significance to determine the harvest time postharvest storage conditions apples. In this study, a portable visible near-infrared (VIS/NIR) analysis device prototype was developed based on multispectral sensor applied ‘Fuji’ apple detection. data apples with variation measured, prediction model established by least-square support vector machine linear discriminant analysis. Due low resolution data, regular preprocessing methods cannot improve accuracy. Instead, spectral shape features (spectral ratio, difference, normalized intensity difference) were used for establishment, combination three effectively improved performance accuracy 88.46%. addition, validation optimal 84.72%, area under curve (AUC) value each level higher than 0.8972. results show that an appliable choice development maturity, processing method proposed in study provides potential solution sensors.