Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Abstract. Near infrared reflectance spectroscopy (NIRS) is a new technology which has been widely used due to the advantages of rapid, nondestructive detection of multiple components content in the same time. In this paper, the Bluetooth-connected DLP mobile NIR spectrometer for portable chemical analysis (DLP® NIRscan^TM Nano Evaluation Module) produced by Texas Instruments was used to identify the different varieties of corn kernels harvested in the same year. The instrument has the advantages of compact size, high signal-to-noise ratio, self-configuring scanning configuration. Savitzky-Golay (S-G) smoothing, multivariate scatter correction (MSC) and Standard normal variate (SNV) were applied on the collected spectral data to eliminate spectral noise and remove the singular spectrum. Then, the principal component analysis (PCA) method was used to transform data and reduce the dimensionality. Finally, the support vector machine (SVM) algorithm was used to construct the classification models of three varieties of corn kernels, and the best classification accuracy was up to 80%. The results showed that the mobile NIR spectrometer combined with chemometric methods could be used for the development of online or portable devices for corn variety identification.