Researchers from the Hefei Institutes of Physical Science (HFIPS) of the Chinese Academy of Sciences have proposed a novel analytical method, based on near infrared (NIR) spectroscopy and data fusion. They show that there is improvement in the accuracy of the spectrometric method for detecting the quality of agricultural products.
The new method is developed based on the fusion of the NIR spectral signals measured in diffuse reflectance (NIRr) and diffuse transmission (NIRt). The researchers assumed that the NIRr and NIRt spectra of the same set of samples are complementary, so the fusion of the two types of spectral signals can provide more complete sample information. By analysing the NIRr and NIRt spectra of three groups of rice flour samples and selecting appropriate chemometric algorithms to extract and integrate the complementary information, the researchers established multiple calibration models to achieve more accurate predictions of the three main components (including amylose, protein and fat content) of rice flour.
This method could also help seed breeders select high-quality rice varieties and help grain producers produce better-quality rice more effectively. It could also be applied to quality detection of other products in the future.