Food Quality and Food Safety
In industrial food processing, quality control often remains a compromise between effort and informational value. Despite significant investments in manual testing methods, sampling, and visual imaging systems, inadequately captured quality attributes lead to food losses amounting to billions every year. What is missing is an objective, scalable, and fully data-driven measurement method—one that goes beyond mere visual inspections and makes the actual product quality measurable.
Hyperspectral imaging addresses this exact need. This technology expands the traditional concept of a camera by adding a dimension: each pixel contains not only color or brightness information but a complete light spectrum.
From this, characteristic spectral signatures arise for each object, allowing precise conclusions about chemical, biological, and physical composition—non-destructive, contactless, and in real time.
Precision in Real Time – Along the Production Line
Spectral analysis enables the food industry’s production process to infer parameters such as fat, protein, and water content—without direct contact and in real time. For meat products, tenderness and texture can also be analyzed, while molecular signatures allow objective quality assessments. Hyperspectral cameras detect differences in the spectral behavior of various materials, enabling the detection of foreign objects like bones, cartilage, and tendons, as well as plastic, wood, or metal particles.
Biological contaminants such as parasites, often missed by conventional methods, are also identified.
In fruit and vegetable processing, hyperspectral analysis reveals cosmetic defects and bruises hidden beneath the skin. It precisely measures ripeness and biochemical quality, allowing exact determination of the optimal harvest time. Seal inspections on packaging especially benefit from the ability to detect micro-contaminations between sealing layers and to identify contamination through transparent or printed films.
In food sorting, the technology ensures the highest standards of food safety and minimizes waste. It sorts dry products such as seeds, rice, nuts, grains, legumes, and spices with utmost precision. For animal products, it analyzes meat, poultry, seafood, and cheese for quality parameters. The imaging is performed at multiple stages along the production line—from raw material inspection through intermediate processes to final quality control before packaging.
New Perspectives for Control and Automation
In the industrial context, this means that the ripeness of fruit, the cell structure of fresh vegetables, the water content of leafy greens, or early microbial changes can be analyzed immediately upon receipt or during the ongoing production process—without sample preparation and without delay. The spectral information replaces the subjective assessment by the operator. The result is a new, reproducible objectivity in quality control that can be fully automated.
As part of an industry-oriented research project, a hyperspectral snapshot system was developed specifically tailored to the requirements of food processing. It combines robust hyperspectral cameras with stable LED lighting and software integration for machine-learning-supported image processing. This system was successfully tested on a wide variety of fruits and vegetables—including sensitive products like berries, stone fruits, and leafy greens. It demonstrated that spectral signatures can detect differences in ripeness, microbial contamination, bruises, or dehydration processes significantly earlier than visual or tactile methods.
The real added value, however, arises from combining these spectral data with data-driven classification and prediction models. With the help of trainable algorithms, shelf life can be predicted, quality losses during transport identified, or automated sorting processes realized based on actual content rather than just optical appearance. Moreover, integration into cloud-based systems enables linking with external data sources—such as weather data from the harvest location or traceability information—thus creating, for the first time, a fully digital quality profile along the entire supply chain.
Understanding Quality — Instead of Just Seeing
For industrial practice, this opens up new perspectives: quality control is no longer seen as a cost factor but as a data-driven control variable that reduces waste, minimizes recall risks, and transparently documents product quality. In research, it enables new approaches for precise analysis, automated classification, and real-time monitoring of biological systems under production conditions.
Hyperspectral imaging thus becomes a central component of the digital fresh economy—serving as an interface between optical metrology, AI-supported analysis, and industrial automation. It replaces mere seeing with precise understanding of complex product states—and thereby lays the foundation for a new quality of quality.
About the Author
Dr. Matthias Locherer has been the Sales Director at Cubert GmbH since 2017. With a PhD in Earth Observation from Ludwig Maximilian University of Munich, he brings extensive expertise in remote sensing, spectral imaging, and data analysis. Matthias has contributed to numerous research projects and publications, particularly in the hyperspectral monitoring of biophysical and biochemical parameters using hyperspectral satellite missions. His deep knowledge of optical measurement techniques and physical modeling makes him a key driver in advancing innovative hyperspectral technologies at Cubert.
