FTNIR-A Robust Diagnostic Tool for the Rapid Detection of Rhyzopertha dominica and Sitophilus oryzae Infestation and Quality Changes in Stored Rice Grains

• Published in: Food and bioprocess technology 2018-04, Vol.11 (4), p.785-796
• Main Authors: Srivastava, Shubhangi, Mishra, Gayatri, Mishra, Hari Niwas
• Format: Article
• Language: English
• Subjects: Agriculture; Algorithms; Biotechnology; Chemistry; Chemistry and Materials Science; Chemistry/Food Science; Computer applications; Diagnostic software; Diagnostic systems; Food Science; Fourier transforms; Infestation; Infrared spectroscopy; Original Paper; Parameter estimation; Physicochemical properties; Quality management; Regression analysis; Rhyzopertha dominica; Rice; Robustness (mathematics); Sitophilus oryzae; Spectrum analysis
• ISSN: 1935-5130; 1935-5149
• DOI: 10.1007/s11947-017-2048-3

Fourier transform infrared spectroscopy (FTNIR) is an excellent mode for evaluation of grain-quality attributes. It enables the nonperturbative molecular information to be diagnosed and allows the explication of images of grains by the passage of the spectral data through an array of computational algorithms. The images are contrived from fingerprint spectra so the conception is that the reflection can conceal the status of the analyzed sample. Rhyzopertha dominica F.- and Sitophilus oryzae- infested and fresh rice grains analyzed with FTNIR within a range of 12,000–4000 cm −1 were proffered to mathematical processing. Partial least squares regression (PLSR) was used for the estimation of physicochemical parameters of rice grains. Outstanding predictive results were acquired denoting that infested rice grains could be convincingly quantified. The coefficient of correlation, root mean square error of validation, and cross validation for the FTNIR model developed to quantify the quality attribute changes with infestation in rice grains were in the range of 99.85–99.01% ( R 2 ), 0.2–1.14% (RMSEE), and 0.3–1.25% (RMSECV). Excellent prediction results of various physico-chemical attributes were obtained for rice grains indicating the fresh and infested samples can be uniquely identified. Also, a paired t test was performed to compare the analytical methods with FTNIR-developed method; no significant difference was found ( t cal 0.025  6) between the two methods. Thus, the results further confirmed the developed FTNIR system to be inherently rapid, clean, and capable of preventing hazardous chemicals which originates from traditional analytical processes and has the potential for monitoring and sorting of rice grains.