Vision inspection systems and the quality and safety of pet food products

Animal food production takes place under dynamic production conditions. High production line speeds, large batch volumes or an environment conducive to microbial growth make random manual inspection insufficient. Also, the quality requirements of retail chains and consumer expectations mean that the pressure for reproducible, stable food quality is increasing. Is automated vision inspection of pet food the answer to these challenges?

Advantages of vision inspection over manual inspection

Limitations of manual control

• The quality of the inspection depends on the employee’s well-being – fatigue, boredom or ill health can negatively affect its reliability.
• The products to be checked move more slowly so that a person is able to assess the condition of the required number of products.
• People, especially in the case of subjective evaluation, can make mistakes, and the process is not repeatable.
• The inspections are random.
• The data is entered manually into the system for later analysis.

What does video inspection enable?

Automatic inspection using machine vision, on the other hand, makes the decision to discard a defective product based on objective software evaluation. In addition, it classifies images at a much higher speed than a human. In the case of can inspection, this is as high as 300 pieces per minute. Consequently, the system inspects every product, and the collected data is recorded automatically and analyzed in real time. This makes it possible to react quickly if a batch defect is detected.

Hyperspectral technology detects foreign bodies

One of the key safety considerations that pet food must meet is the absence of foreign bodies. If a piece of plastic or metal is eaten, it can cause poisoning of the animal or serious damage to its health. Such events negatively affect the brand image.

Using hyperspectral technology and machine learning (ML) algorithms, even small foreign bodies can be detected. The detection framework is tailored to the context of the production line. Detected objects include plastic, metals, plastics, pieces of feathers and bones, textiles, glass or wood.

Vision inspection is performed for both wet food meat and pellets in the case of dry food.

What defects do vision systems detect?

1. raw material stage

At the raw material inspection stage for animal food, vision inspection detects not only foreign bodies and contaminants. Algorithms also check for color deviations, texture disorders or color uniformity in the raw material.

2nd processing stage

After processing, deviations in shape, color and structural integrity are controlled using machine vision.

3. packaging stage

Vision inspection of the packaging process ensures the freshness and safety of animal food. With its use, the quality of cans, trays or sachets (pouch type packaging) is checked. Particularly in the case of cans, the systems detect a whole range of closing irregularities, such as:

• teeth;
• descent;
• Dents, deformations, and geometric abnormalities;
• creases not connected to the body;
• Excess material protruding beyond the edge;
• Cracks, cuts, and seam jamming;
• cut-off collar.

In addition, printing defects (incomplete printing, printing inconsistent with the batch, inconsistency of the lid with the bottom of the can) and label errors (missing label, peeling or displacement), key defects (missing or displaced) are also detected.

See how vision inspection has increased the quality of cans in the fish industry.

NOK classification – what’s next?

Regardless of the stage of production, in the case of NOK grading, it is possible to integrate a rejector into the control station that automatically removes the product from the production line and logs this event in the database. This allows for:

1. Real-time quality analysis to enable prediction of machine condition and early detection of batch defects
2. quick response of the operator if he receives information about NOK on the operator panel;
3. and rechecking and possible recycling of discarded raw materials.

Effects of system implementation in the pet food industry

• controlling the parameters of pet food products at every stage of the production process, from raw meat to sealed packaging;
• Accelerating and smoothing the production process;
• reducing production losses;
• reduction in the number of complaints;
• ensuring the safety of the feed;
• Stabilizing product quality at a high level;
• Analyzing data from the production line to enable rapid response in case of failure or predictive maintenance;
• Strengthening the brand’s position in the market;
• Increased consumer confidence.