Abstract:In order to meet the requirements of apple internal quality inspection, the apple internal quality sorting manipulator was designed based on the visible/near infrared spectroscopy detection technology and the sorting manipulator. The device consisted of three parts: clamping mechanism, near infrared spectroscopy acquisition system and control system. The manipulator clamped the apple and collected the nearinfrared spectral data of the apple. The spectral data was analyzed by the predictive model in the upper computer software, and the spectral curve and predicted result were displayed. Based on this device, the visible and nearinfrared reflection spectra of apple in the range of 650~1100nm were collected. Totally 200 apples were used for the experiment, including 150 apples in the prediction set and 50 apples in the verification set. The soluble solids content of the apples was measured by temperaturecompensated refractometer after the collection of spectral information. The collected spectra were pretreated by Savitzky-Golay smooth (SG-smooth), standard normal variable transformation (SNV) and multiplication scattering correction (MSC). The partial leastsquares prediction model of the apple’s SSC was established with spectral data as independent variable and soluble solids as dependent variable. The prediction result that preprocessed with the multiscattering correction (MSC) method was the best. The correlation coefficients of the calibration set and the verification set of the prediction model were 0.9782 and 0.9701, respectively, and the root mean square errors were 0.2746°Brix and 0.3263°Brix, respectively. Finally, the accuracy of models was tested. The reflect spectra of 20 samples were collected, and then the real values of these samples’ SSC were measured. The prediction model could give satisfactory results with the correlation coefficient of 0.9573 and the root mean square error of prediction of 0.4224°Brix. The results indicated that this device can satisfy the requirements of apple internal quality detection with high accuracy and good performance.