Modeling for mung bean variety classification using visible and near-infrared hyperspectral imaging

Abstract

This study was carried out to investigate the feasibility of using visible and near infrared hyperspectral imaging for the variety classification of mung beans. Raw hyperspectral images of mung beans were acquired in the wavelengths of 380-1023 nm, and all images were calibrated by the white and dark reference images. The spectral reflectance values were extracted from the region of interest (ROI) of each calibrated hyperspectral image, and then they were treated as the independent variables. The dependent variables of four varieties of mung beans were set as 1, 2, 3 and 4, respectively. The extreme learning machine (ELM) model was established using full spectral wavelengths for classification. Modified gram-schmidt (MGS) method was used to identify effective wavelengths. Based on the selected wavelengths, the ELM and linear discriminant analysis (LDA) models were built. All models performed excellently with the correct classification rates (CCRs) covering 99.17%-99.58% in the training sets and 99.17%-100% in the testing sets. Fifteen wavelengths (432 nm,

Cultivar Classification of Single Sweet Corn Seed Using Fourier Transform Near-Infrared Spectroscopy Combined with Discriminant Analysis

“…Thus, these methods may be used to detect a small group of sampling seeds. The complex sample preparation required for such detection methods also limits the possibilities of using these methods for online detection in the seed industry [10]. In view of these drawbacks, researchers have shown great interest in looking for rapid and non-destructive methods for seed cultivar detection.…”

“…This HSI system was also used to discriminate the varieties of grape seed, where the highest accuracy of 88.7% was achieved in the prediction set [14]. Xie et al, (2018) carried out research to recognize the varieties of mung beans using the HSI (380-1023 nm) system, where the extreme learning machine algorithm performed with accuracies ranging from 99.17% to 100% [10]. utilized two HSI systems with different spectral ranges (380-1030 nm and 874-1734 nm) to identify rice seed varieties.…”

Multi-kernel dictionary learning for maize varieties classification

“…In the domestic research, progress has been made in the detection and identification of maize varieties using machine vision technology. Grains collecting and processing are carried out by machine vision, characteristic parameters of maize are chosen and optimized [4][5][6][7][8][9] by a series of algorithm such as artificial neural network model, support vector machines etc.…”

Application of Visible/Infrared Spectroscopy and Hyperspectral Imaging With Machine Learning Techniques for Identifying Food Varieties and Geographical Origins

“…Summary of selected references for fruits classification with hyperspectral imaging.Based on the spectral and spatial information, HSI could be exploited as a powerful tool for the traceability of black bean (78), honey(122), okra kernels (61), and mung beans(123). It was worth noting that Sun et al(78) combined spectral and image features, and the optimal PLS-DA model obtained the accuracy of 98.33% for classifying black beans from three provinces.…”

“…It was worth noting that Sun et al(78) combined spectral and image features, and the optimal PLS-DA model obtained the accuracy of 98.33% for classifying black beans from three provinces. Also, Xie et al(123) proposed the Modified gram-Schmidt (MGS) method to select effective wavelengths for classification of four mung bean varieties, based on which both ELM and LDA models obtained the prediction accuracy over 98%.…”