Rapid and non-destructive identification of water-injected beef samples using multispectral imaging analysis

Area of Science:

• Food Science
• Analytical Chemistry
• Imaging Technology

Background:

• Water injection in beef is a significant food safety issue, raising public concern.
• Accurate detection methods are needed to ensure meat product integrity and consumer trust.

Purpose of the Study:

• To evaluate multispectral imaging in the visible and near-infrared regions for identifying water-injected beef.
• To develop a quantitative model for estimating actual water increase in beef.

Main Methods:

• Acquisition of multispectral images (405-970 nm) of beef with varying water injection levels.
• Application of partial least squares regression (PLSR) for quantitative prediction of water content.
• Integration of spectral data with RGB image features for model optimization.

Main Results:

• A PLSR model achieved a correlation coefficient (r) of 0.923 for predicting water increase.
• An optimized model combining spectral and RGB data improved prediction accuracy (r = 0.946).
• Visualizations demonstrated the spatial distribution of actual water increase within beef samples.

Conclusions:

• Multispectral imaging is a viable rapid, non-destructive tool for identifying water-injected beef.
• The developed models offer accurate quantitative assessment and spatial mapping of water content.
• This technology can enhance food safety and quality control in the meat industry.