/
Agriculture
Rapid Content Analysis of Meat & Seafood

Meat & Seafood Analysis with Hyperspectral Sensors

Objective, non-invasive and non-destructive techniques for chemical composition analysis, such as hyperspectral imaging, can help operators mitigate risks, reduce costs and unlock value.

About Opsyne

Opsyne builds and sells OEM hyperspectral imaging technology that can be integrated into complete industrial automation solutions.

Recent Advances for Rapid Identification of Chemical Information of Muscle Foods by Hyperspectral Imaging Analysis

Background

The world's average stock of chickens is almost 19 billion, or three per person, according to statistics from the UN's Food and Agriculture Organisation. Cattle are the next most populous breed of farm animal at 1.4 billion, with sheep and pigs not far behind at around 1 billion. Global annual consumer spending on animal meat is $1.4 trillion, or about 15% of total food and grocery.

Muscle foods, including pork, beef, chicken and seafood play provide an essential source of high-quality protein and amino acids. Food safety and quality are crucial. The chemical composition and physical and biological attributes are important information to packers, processors, and retailers.

Traditionally, the chemical analysis of meats is performed by taking samples to a lap. Spot sampling can be time-consuming, destructive, tedious and demanding. It also provides limited visibility. The analysis is intermittent rather than continuous. Reactive than proactive. And only a few samples are inspected rather than a scan of the entire batch or product line. However, objective, non-invasive and non-destructive techniques for chemical composition analysis, such as hyperspectral imaging, can help operators mitigate risks, reduce costs and unlock value. New research has summarized the applications of hyperspectral imaging in this domain.

Review

The chemical composition of muscle foods determines their quality and flavour. Measuring attributes such as protein, fat, salt, moisture content, and pH and freshness attributes with the use of hyperspectral imaging technology are a proven method for ensuring the highest safety and quality of meat and meat products. 

Multiple studies focused on assessing the moisture content of pork and lamb meat using a push-broom HSI system in the NIR range (900-1700 nm) in six different wavelengths. For other seafood products from the likes of prawns and fish fillets, due to their inherent differences, the optimal wavelengths selected and the prediction models used differed but still yielded satisfactory performance in determining moisture content.

Fat and fatty acid content are two important chemical components that impact the cooked properties of meat, such as flavour and tenderness, specifically intramuscular fat, which is usually subjectively assessed based on marbling. Multiple studies showed promise in predicting the fat content in pork and beef. In addition to beef and pork, fat prediction associated with hyperspectral was carried out in seafood products and showed satisfactory performance in wavelength regions ranging from 760-1649 nm. 

In addition to moisture and fat, protein is a significant meat component and is of high nutritional value. Its accurate measurement is also crucial due to its changing properties during storage from microorganisms and enzymes that can degrade the nutrition and economic benefit. Hyperspectral can monitor meat products throughout processing as well as the supply chain.

The studies of fat and moisture also assessed pork, lamb, and beef's protein content, all showing similar satisfactory results. In fresh meat, protein content in cooked hams has been analyzed with a hyperspectral technique in the near-infrared (NIR) spectral region and indicated the ability to predict the protein contents in cooked hams and poultry meats. Predicting individual amino acid content in chicken has been demonstrated as well.

Implications

Hyperspectral can quantify and map the chemical composition of muscle foods. The non-invasive, non-destructive and rapid technology can be implemented for industrial application. The latest advancements in optics and machine learning are enabling industrial-grade performance.

Future studies determining the optimal wavelengths containing the most valuable information with more advanced algorithms, potentially creating an automatic selection, could lower the redundant data collected, which would further improve the quality of prediction while reducing the computational time and cost. Continued application of hyperspectral imaging throughout the food industry is expected as the development of computing and chemometric techniques progresses to improve rapid detection and real-time information collection. 

Application notes