Abstract:Journal of Near Infrared Spectroscopy, Ahead of Print. Sorghum has a long history of cultivation and is an important food and economic crop. It can be divided into glutinous and non-glutinous varieties according to the starch structure and content. Rapid discrimination between the two would help the winemaking, feed, and food industries complete purchase pricing, ingredients, and quality control. In this study, 38 different samples were acquired, including 14 glutinous and 24 non-glutinous sorghum samples. Near infrared (NIR) spectra of glutinous and non-glutinous sorghum, pre-treated using the standard normal variable (SNV) transformation were found to have slightly different absorbances in the combination and first overtone bands. Based on the distribution of the starch-related and hydrogen-containing groups in the NIR region, it was concluded that glutinous sorghum has more C-O and C-C groups than non-glutinous sorghum. This study proposes an approach based on typical samples and direct calibration (TSDC) for binary discrimination. The TSDC approach consists of three functions. First, typical samples of two types of samples were selected. Second, typical type samples are used as dependent variables, predicted samples are used as independent variables, and formula regression is used to obtain fitted coefficients. Finally, if the formula regression model has no solution or the fitted coefficient is 1, typical type samples are reselected. Using the TSDC approach, discrimination accuracy can achieve 100% accuracy at 0.5 threshold. A larger threshold can be set to select better type characteristic predicted samples for discrimination. The TSDC approach can build excellent model through real relevance between the NIR spectra and the properties of interest, and the use of typical type samples greatly reduces modeling work compared with complex pattern recognition methods, especially for highly varied agricultural products. Therefore, it can efficiently propel the application and development of NIR detection technology. More research is required to apply the TSDC approach to three types of samples.

Accurate determination of sorghum origin based on hyperspectral imaging technology and machine learning

Application of Hyperspectral Reflectance Imaging and Chemometric Methods to Classify the Corn Seed Variety

Classification of Pesticide Residues in Sorghum Based on Hyperspectral and Gradient Boosting Decision Trees

Classification and characterization on sorghums based on HS-GC-IMS combined with OPLS-DA and GA-PLS

The Rapid Non-Destructive Detection of the Protein and Fat Contents of Sorghum Based on Hyperspectral Imaging

A Portable Method for Predicting the Germination Percentage of Sorghum × Sudangrass Seed Using Multispectral Images

Combination of the Manifold Dimensionality Reduction Methods with Least Squares Support Vector Machines for Classifying the Species of Sorghum Seeds

Hyperspectral imaging combined with GA-SVM for maize variety identification

Smartagb: Aboveground Biomass Estimation of Sorghum Based on Spatial Resolution, Machine Learning and Vegetation Index

Combined laser-induced breakdown spectroscopy and hyperspectral imaging with machine learning for the classification and identification of rice geographical origin

Detecting Sorghum Plant and Head Features from Multispectral UAV Imagery

Application of Hyperspectral Imaging to Discriminate the Variety of Maize Seeds

Aerial Imagery Analysis – Quantifying Appearance and Number of Sorghum Heads for Applications in Breeding and Agronomy

Spectral and Image Integrated Analysis of Hyperspectral Data for Waxy Corn Seed Variety Classification

Application of Hyperspectral Imaging Technology to Discriminate Different Geographical Origins of Jatropha Curcas L. Seeds

Near infrared spectroscopy discriminates glutinous and non-glutinous sorghum using an approach based on typical samples and direct calibration

Application Of Hyperspectral Imaging And Chemometric Calibrations For Variety Discrimination Of Maize Seeds

Study on Method of Maize Hybrid Purity Identification Based on Hyperspectral Image Technology

Nondestructive Detection of Sunflower Seed Vigor and Moisture Content Based on Hyperspectral Imaging and Chemometrics

Hyperspectral Imaging with Machine Learning for Non-Destructive Classification of Astragalus Membranaceus Var. Mongholicus, Astragalus Membranaceus, and Similar Seeds.

Sorghum Grains Grading for Food, Feed, and Fuel Using NIR Spectroscopy