Non-destructive prediction of some quality characteristics of orange fruit using Vis-NIR spectroscopy

Due to the economic value and the high cost of various inputs used to produce agricultural products, measuring their quality characteristics using non-destructive methods, is an important requirement. Among various non-destructive techniques, optical spectroscopy methods are more important because they are rapid and easy to use. In this research, visible/near-infrared (Vis-NIR) (400-900 nm) spectroscopy was used to predict the quality characteristics of Thomson oranges. Some physicochemical properties of oranges including dimensions, weight, volume weight, soluble solids content (SSC), acidity (TA), pH, as well as SSC/TA and BrimA indices were determined. Vis-NIR spectra were pre-processed using different smoothing and normalization methods and their combinations. The results showed that the Vis-NIR spectroscopy with median filter and multiplicative scatter correction (MSC) pre-processing methods can well predict SSC (rp = 0.983) and pH (rp = 0.972) of Thomson oranges. Moreover, this technique combined with median filter and MSC methods could predict TA (rp = 0.771) and BrimA index (rp = 0.865) of Thomson oranges with acceptable results. The SSC/TA index was also predicted with acceptable results using Vis-NIR spectroscopy combined with the median filter and MSC methods (rp = 0.829). Therefore, Vis-NIR spectroscopy as a non-destructive technique can be used for quality evaluation of Thomson oranges.