Where does efficiency end and quality risk begin? About vision systems in the fish industry.

The fish industry in Poland today operates under conditions of increasing production scale and high quality requirements imposed by both the market and food safety systems. In 2024, about 362,000 tons of fish and fish products with a total value of about EUR 3.1 billion were exported from Poland. Such a high volume means that even single quality errors can result in significant financial losses or image risks.

Production efficiency vs. stability of quality control

As production lines increase in speed, the difficulty of maintaining stable and repeatable quality control increases. Manual inspection, based on the visual judgment of operators, is naturally susceptible to fatigue, variability of attention and time pressure. To illustrate the scale: with a line operating at 300 pcs/min, even a 3% oversight rate means about 9 potentially misjudged products per minute (illustrative example). In continuous production, this level of risk becomes difficult to accept from a food safety perspective.

Automatic quality control of raw fish and cans

Scanway vision systems used in the fish industry realize automatic quality control of both raw fish meat and cans (round, Club or Hansa), continuously, without affecting the productivity of production lines, while meeting IP67 food industry standards. Both fish meat and cans are evaluated objectively, based on real-time image analysis. Data processing is carried out using Machine Learning/ Deep Learning algorithms, which are responsible for classifying nonconformities and deciding whether to reject the product.

Quality control of fish meat – scope and technology

For fish meat inspection, Scanway’s systems use RGB and hyperspectral (HSI) cameras, which allows simultaneous analysis of visual features and material properties of the raw material. This approach makes it possible to detect quality defects as well as foreign bodies made of plastic or biological elements, even when they are poorly visible in standard visual inspection.

The system implements detection of foreign bodies, such as:

• High and low density plastics (PET, PE, PP),

• fragments of gloves, rubber, textiles and films (including transparent ones),

• wood, glass, paper,

• Biological elements, including bones.

The minimum detectable size of foreign bodies is 10 × 10 mm (HSI / RGB+HSI). At the same time, meat quality defects are identified, including blood spots (5 × 5 mm and above), discoloration, skin residues, delamination of the muscle structure and missing shell fragments. The efficiency of detecting quality defects reaches up to 95%. Each piece of raw material is automatically classified as OK or NOK, and product that does not meet quality criteria – thanks to integration with an automatic rejector – can be physically removed from the line in real time.

Inspection of cans as a critical element of safety

In canned fish production, product safety depends not only on the quality of the raw material, but also on the integrity of the packaging. Scanway’s vision systems perform continuous can inspection at full line speed, with no product contact, at up to 300 cans/min.

The automated inspection includes, among other things:

• Seam defects: teeth, drooping, dents, creases not connected to the body, excess material protruding beyond the edge, cracks and cuts in the seam, cut flange, jammed seams,

• disturbed can geometry, including dents and deformations,

• Defective printing: incomplete printing, printing inconsistent with the batch, inconsistency of the lid with the bottom of the can,

• Label errors: the label is missing, unstuck or misplaced,

• Key defects: the key is missing or misplaced.

Each can is analyzed individually, and a decision on a possible rejection is made automatically in a fraction of a second. As a result, the inspection of cans ceases to be random and becomes a permanent part of the production safety system.

READ MORE: can defects detected by vision system

Automation as a response to industry challenges

Vision systems address the key challenges of today’s fishing industry: lack of manpower for manual inspection, short production runs and frequent changeovers, losses due to quality errors, and the need for data collection and analysis as part of process digitization. Automated quality control provides repeatable decisions, full traceability and better quality risk management.

Under conditions of increasing efficiency, safety in the fish industry is no longer a single inspection step. It is becoming a continuous process whose stability is increasingly based on automatic, objective inspection of both raw material and cans.