AI-Powered Smart Food Drying Techniques Enhance Product Quality and Efficiency

Smart food drying techniques with AI enhance product quality and efficiency

by Marianne Stein, College of Agricultural, Consumer and Environmental Sciences at the University of Illinois Urbana-Champaign Food drying is a common process for preserving many types of food, including fruits and meat; however, drying can alter the food's quality and nutritional value. In recent years, researchers have developed precision techniques that use optical sensors and AI to facilitate more efficient drying. A new study from the University of Illinois Urbana-Champaign discusses three emerging smart drying techniques, providing practical information for the food industry. The paper is published in the journal Food Engineering Reviews. "With traditional drying systems, you need to remove samples to monitor the process. But with smart drying, or precision drying, you can continuously monitor the process in real time, enhancing accuracy and efficiency," said corresponding author Mohammed Kamruzzaman, assistant professor in the Department of Agricultural and Biological Engineering (ABE), part of the College of Agricultural, Consumer and Environmental Sciences and The Grainger College of Engineering at Illinois. In the paper, the researchers review academic literature about different types of equipment that apply precision techniques to enhance smart drying capabilities in the food industry. They focus on three optical sensing systems -- RGB imaging with computer vision, near-infrared (NIR) spectroscopy, and near-infrared hyperspectral imaging (NIR-HSI) -- discussing the mechanisms, applications, advantages, and limitations of each. They also provide an overview of standard industrial drying methods, such as freeze drying, spray, microwave, or hot-air oven drying, which can be combined with precision monitoring techniques. "You can use each of the three sensors separately or in combination. What you choose will depend on the particular drying system, your needs, and cost-effectiveness," said lead author Marcus Vinicius da Silva Ferreira, a postdoctoral fellow in ABE. RGB with computer vision uses a regular camera that captures visible light with an RGB color spectrum. It can provide information about surface-level features, such as size, shape, color, and defects, but it is not capable of measuring moisture content. NIR spectroscopy uses near-infrared light to measure the absorbance of different wavelengths, which can be correlated to unique chemical and physical product characteristics, and it can measure internal qualities such as moisture content. However, NIR scans one point at a time. This can work for a single product, like an apple slice, at least initially, Kamruzzaman said. "But as the drying progresses, the material will shrink and become heterogeneous, because of cracking and bending. If you use NIR at that stage, and if you only scan a single point, you cannot measure the drying rate," he noted. NIR-HSI is the most comprehensive of the three techniques. It scans the whole surface of the product, so it provides much more precise information about the drying rate and other features than NIR alone, since it extracts three-dimensional spatial and spectral information. However, NIR-HSI is also much more expensive than the two other sensors. The equipment costs 10 to 20 times more than NIR sensors, and 100 times or more than RGB cameras. Additionally, maintenance and computing requirements for HSI are substantially higher, further increasing the total cost. All three methodologies must be combined with AI and machine learning to process the information, and the models must be trained for each specific application. Again, HSI requires more computational power than the other two systems because of the large amount of data it collects. The researchers also developed their own drying system to test the various methods. They built a convective heat oven and tested the techniques on the drying of apple slices. They first combined the system with RGB and NIR; later they also tested the NIR-HSI system, the findings of which they plan to discuss in a forthcoming paper. "For real-time monitoring, the convergence of RGB imaging, NIR spectroscopic sensors, and NIR-HSI with AI represents a transformative future for food drying. Integrating these technologies overcomes conventional drying process monitoring limitations and propels real-time monitoring capabilities," they concluded in the paper. Future development of portable, hand-held NIR-HSI devices will further enable continuous monitoring of drying systems, providing real-time quality control in a variety of operating environments, they noted.

Smart food drying techniques with AI enhance product quality and efficiency

Food drying is a common process for preserving many types of food, including fruits and meat; however, drying can alter the food's quality and nutritional value. In recent years, researchers have developed precision techniques that use optical sensors and AI to facilitate more efficient drying. A new study from the University of Illinois Urbana-Champaign discusses three emerging smart drying techniques, providing practical information for the food industry. "With traditional drying systems, you need to remove samples to monitor the process. But with smart drying, or precision drying, you can continuously monitor the process in real time, enhancing accuracy and efficiency," said corresponding author Mohammed Kamruzzaman, assistant professor in the Department of Agricultural and Biological Engineering (ABE), part of the College of Agricultural, Consumer and Environmental Sciences and The Grainger College of Engineering at Illinois. In the paper, the researchers review academic literature about different types of equipment that apply precision techniques to enhance smart drying capabilities in the food industry. They focus on three optical sensing systems -- RGB imaging with computer vision, near-infrared (NIR) spectroscopy, and near-infrared hyperspectral imaging (NIR-HSI) -- discussing the mechanisms, applications, advantages, and limitations of each. They also provide an overview of standard industrial drying methods, such as freeze drying, spray, microwave, or hot-air oven drying, which can be combined with the precision monitoring techniques. "You can use each of the three sensors separately or in combination. What you choose will depend on the particular drying system, your needs, and cost-effectiveness," said lead author Marcus Vinicius da Silva Ferreira, a postdoctoral fellow in ABE. RGB with computer vision uses a regular camera that captures visible light with a RGB color spectrum. It can provide information about surface-level features, such as size, shape, color, and defects, but it is not capable of measuring moisture content. NIR spectroscopy uses near-infrared light to measure the absorbance of different wavelengths, which can be correlated to unique chemical and physical product characteristics, and it can measure internal qualities such as moisture content. However, NIR scans one point at a time. This can work for a single product, like an apple slice, at least initially, Kamruzzaman said. "But as the drying progresses, the material will shrink and become heterogeneous, because of cracking and bending. If you use NIR at that stage, and if you only scan a single point, you cannot measure the drying rate," he noted. NIR-HSI is the most comprehensive of the three techniques. It scans the whole surface of the product, so it provides much more precise information about the drying rate and other features than NIR alone, since it extracts three-dimensional spatial and spectral information. However, NIR-HSI is also much more expensive than the two other sensors. The equipment costs 10 to 20 times more than NIR sensors, and 100 times or more than RGB cameras. Additionally, maintenance and computing requirements for HSI are substantially higher, further increasing the total cost. All three methodologies must be combined with AI and machine learning to process the information, and the models must be trained for each specific application. Again, HSI requires more computational power than the other two systems because of the large amount of data it collects. The researchers also developed their own drying system to test the various methods. They built a convective heat oven and tested the techniques on the drying of apple slices. They first combined the system with RGB and NIR; later they also tested the NIR-HSI system, the findings of which they plan to discuss in a forthcoming paper. "For real-time monitoring, the convergence of RGB imaging, NIR spectroscopic sensors, and NIR-HSI with AI represents a transformative future for food drying. Integrating these technologies overcomes conventional drying process monitoring limitations and propels real-time monitoring capabilities," they concluded in the paper. Future development of portable, hand-held NIR-HSI devices will further enable continuous monitoring of drying systems, providing real-time quality control in a variety of operating environments, they noted.

Smart food drying techniques with AI enhance produ | Newswise

URBANA, Ill. - Food drying is a common process for preserving many types of food, including fruits and meat; however, drying can alter the food's quality and nutritional value. In recent years, researchers have developed precision techniques that use optical sensors and AI to facilitate more efficient drying. A new study from the University of Illinois Urbana-Champaign discusses three emerging smart drying techniques, providing practical information for the food industry. "With traditional drying systems, you need to remove samples to monitor the process. But with smart drying, or precision drying, you can continuously monitor the process in real time, enhancing accuracy and efficiency," said corresponding author Mohammed Kamruzzaman, assistant professor in the Department of Agricultural and Biological Engineering (ABE), part of the College of Agricultural, Consumer and Environmental Sciences and The Grainger College of Engineering at Illinois. In the paper, the researchers review academic literature about different types of equipment that apply precision techniques to enhance smart drying capabilities in the food industry. They focus on three optical sensing systems - RGB imaging with computer vision, near-infrared (NIR) spectroscopy, and near-infrared hyperspectral imaging (NIR-HSI) - discussing the mechanisms, applications, advantages, and limitations of each. They also provide an overview of standard industrial drying methods, such as freeze drying, spray, microwave, or hot-air oven drying, which can be combined with the precision monitoring techniques. "You can use each of the three sensors separately or in combination. What you choose will depend on the particular drying system, your needs, and cost-effectiveness," said lead author Marcus Vinicius da Silva Ferreira, a postdoctoral fellow in ABE. RGB with computer vision uses a regular camera that captures visible light with a RGB color spectrum. It can provide information about surface-level features, such as size, shape, color, and defects, but it is not capable of measuring moisture content. NIR spectroscopy uses near-infrared light to measure the absorbance of different wavelengths, which can be correlated to unique chemical and physical product characteristics, and it can measure internal qualities such as moisture content. However, NIR scans one point at a time. This can work for a single product, like an apple slice, at least initially, Kamruzzaman said. "But as the drying progresses, the material will shrink and become heterogeneous, because of cracking and bending. If you use NIR at that stage, and if you only scan a single point, you cannot measure the drying rate," he noted. NIR-HSI is the most comprehensive of the three techniques. It scans the whole surface of the product, so it provides much more precise information about the drying rate and other features than NIR alone, since it extracts three-dimensional spatial and spectral information. However, NIR-HSI is also much more expensive than the two other sensors. The equipment costs 10 to 20 times more than NIR sensors, and 100 times or more than RGB cameras. Additionally, maintenance and computing requirements for HSI are substantially higher, further increasing the total cost. All three methodologies must be combined with AI and machine learning to process the information, and the models must be trained for each specific application. Again, HSI requires more computational power than the other two systems because of the large amount of data it collects. The researchers also developed their own drying system to test the various methods. They built a convective heat oven and tested the techniques on the drying of apple slices. They first combined the system with RGB and NIR; later they also tested the NIR-HSI system, the findings of which they plan to discuss in a forthcoming paper. "For real-time monitoring, the convergence of RGB imaging, NIR spectroscopic sensors, and NIR-HSI with AI represents a transformative future for food drying. Integrating these technologies overcomes conventional drying process moni­toring limitations and propels real-time monitoring capa­bilities," they concluded in the paper. Future development of portable, hand-held NIR-HSI devices will further enable continuous monitoring of drying systems, providing real-time quality control in a variety of operating environments, they noted. The paper, "AI-Enabled Optical Sensing for Smart and Precision Food Drying: Techniques, Applications and Future Directions," is published in Food Engineering Reviews [DOI: 10.1007/s12393-024-09388-0]. This study was financially supported by the Cen­ter for Advanced Research in Drying (CARD), a U.S. National Science Foundation Industry University Cooperative Research Center. CARD is located at Worcester Polytechnic Institute and the University of Illinois at Urbana-Champaign.