Development of Portable Near Infrared Spectroscopy Device and Its Application in Rapid Nondestructive Detection of Strawberry Quality

Abstract:
In order to meet the demand for rapid detection of strawberry quality, we have developed a kind of portable near-infrared spectrometer. Near-infrared spectrograms of "Tao Xun" and "Hong Yan" strawberries were obtained in the range of 600~1100 nm, and the collected spectra were pretreated by multiplication scattering correction (MSC), standard normal variable transformation (SNV), savitzky-golay derivative (SG-DEV), savitzky-golay smooth (SG-smooth), normalization (NOR). Finally, multiple quality prediction models of strawberries were established with principal component analysis (PCA). The results showed that the pre-processed SSC prediction model of MSC was optimal, with high correlation coefficient (Rc = 0.91, Rp = 0.88) and small root mean square error (RMSEC = 0.43%, RMSEP = 0.45%); the preprocessed firmness prediction model of SG-DEV (width 5 points, fitting 3 times) was optimal, with high correlation coefficient (Rc = 0.92, Rp = 0.93) and small root mean square error (RMSEC = 1.20 N, RMSEP = 1.39N). Based on the WINCE development framework, a real-time analysis and control software for nondestructive testing was written, and predictive models of two varieties were added to the software to realize one-click operation of strawberry multi-quality parameter testing. In addition, the prediction model of strawberry quality was verified. The correlation coefficients of SSC model and firmness model were 0.93 and 0.95, respectively, and the root mean square error was 0.44% and 1.15 N, respectively. Compared with traditional detection methods, it is fast and non-destructive, and is suitable for immediate detection by fruit farmers after picking. The portable near-infrared spectrometer is small, light, and inexpensive, which provides a reference for application of the product.