Spectroscopy and computer vision techniques for noninvasive analysis of legumes: A review

Legumes play a vital role in human consumption, animal foraging, and sustainable farming. Rapid quality monitoring techniques in the periods of planting, harvesting, and storage are the major concerns of consumers and food processors. Noninvasive techniques in terms of spectroscopy and computer vision (CV) have become promising in the rapid assessment of the quality and safety of legumes. This review provides a comprehensive overview of detailed applications for determining the quality and safety attributes of legumes using spectroscopy and CV techniques. Spectroscopy approaches include visible and near-infrared (Vis-NIR), near-infrared (NIR), mid-infrared (MIR), Raman, fluorescence, and terahertz (THz) spectroscopy. CV methods include traditional CV, hyperspectral imaging (HSI), and multispectral imaging (MSI). Spectroscopy and CV technologies both possess the advantages of rapid and noninvasive analysis. They can meet the diverse detection requirements of legumes when combined with appropriate data analysis methods. However, there are some challenges with regard to spectral correction algorithms, hardware and instrument development, and image processing techniques that need to be overcome. Much more work is required to develop unified calibration models and specialized equipment before application in the agriculture industry.

Automated summary

Legumes have significant importance in human consumption, animal foraging, and sustainable farming. Spectroscopy and computer vision techniques offer rapid and noninvasive analysis for assessing the quality and safety of legumes. Spectroscopy approaches include visible and near-infrared, near-infrared, mid-infrared, Raman, fluorescence, and terahertz spectroscopy, while computer vision methods include traditional CV, hyperspectral imaging, and multispectral imaging. However, challenges regarding spectral correction algorithms, hardware and instrument development, and image processing techniques need to be addressed for wider application in the agriculture industry.