Food drying is a widely used technique to preserve various food items, such as fruits and meats. However, this process can change the food’s quality and nutritional content. Recently, researchers have come up with advanced methods utilizing optical sensors and artificial intelligence (AI) to improve the drying efficiency. A new study explores three innovative smart drying technologies and offers useful insights for the food industry.
Food drying is a widely utilized method to preserve different food items, including fruits and meats. Nonetheless, this method can modify the food’s quality and nutritional value. In recent times, researchers have introduced advanced techniques that use optical sensors and AI for enhanced drying efficiency. A recent study from the University of Illinois Urbana-Champaign examines three innovative smart drying techniques, presenting valuable information for the food sector.
“Traditional drying methods require removing samples for process monitoring. In contrast, smart or precision drying enables real-time monitoring of the process, improving accuracy and efficiency,” explained Mohammed Kamruzzaman, the study’s lead author and assistant professor in the Department of Agricultural and Biological Engineering (ABE) at the College of Agricultural, Consumer and Environmental Sciences and The Grainger College of Engineering at Illinois.
The paper reviews existing academic research on various types of equipment applying precision techniques to improve smart drying capabilities within the food industry.
The focus is on three optical sensing technologies — RGB imaging with computer vision, near-infrared (NIR) spectroscopy, and near-infrared hyperspectral imaging (NIR-HSI). The researchers discuss how each works, its uses, benefits, and drawbacks. They also examine standard drying methods used in industry, such as freeze drying, spray drying, microwave drying, and hot-air oven drying, which can be paired with precision monitoring tools.
“You can choose to use each of the three sensing technologies individually or in combination, depending on your specific drying needs and cost considerations,” noted Marcus Vinicius da Silva Ferreira, the lead author and a postdoctoral fellow in ABE.
RGB imaging with computer vision employs a standard camera that records visible light across the RGB color spectrum. It can give insights into surface features like size, shape, color, and defects, but it cannot assess moisture levels.
NIR spectroscopy utilizes near-infrared light to analyze the absorbance at various wavelengths. This can be correlated to distinct chemical and physical characteristics of the product, including moisture levels. However, NIR measures only one point at a time.
Initially, this is effective for drying individual items, like an apple slice, according to Kamruzzaman.
“As drying continues, the material tends to shrink and may become uneven due to cracking and bending. If you only perform a single point scan with NIR during this phase, it won’t accurately reflect the drying rate,” he explained.
NIR-HSI offers the most extensive capabilities among the three techniques. It scans the entire surface of a product, yielding much more accurate details regarding the drying rate and additional characteristics than NIR alone, as it gathers 3D spatial and spectral data. However, NIR-HSI comes at a significantly higher price point, costing 10 to 20 times more than NIR sensors and over 100 times that of RGB cameras. Additionally, it demands more maintenance and computing resources, further increasing its overall cost.
All three methods require integration with AI and machine learning for effective data processing, and the models must be tailored for each specific use. Moreover, HSI requires greater computational capacity due to the vast data it gathers.
The team also created their own drying system to test these various methods. They constructed a convective heat oven and evaluated the drying of apple slices using RGB and NIR, before subsequently testing the NIR-HSI system, the results of which will be detailed in a future paper.
“The integration of RGB imaging, NIR spectroscopic sensors, and NIR-HSI with AI for real-time monitoring indicates a groundbreaking future for food drying. Combining these technologies helps overcome the limitations of traditional drying process monitoring and enhances real-time capabilities,” the researchers concluded in the paper.
Future advancements may lead to portable, handheld NIR-HSI devices, enabling ongoing monitoring of drying systems and providing real-time quality control across various environments, as noted by the researchers.
