Automation of packaging inspection throughout the manufacturing and distribution process
Reduce unnecessary rework with accurate inspection
Ensuring consistent product quality


Defective or damaged packaging can negatively affect distributor and consumer perceptions of product quality, safety, and value, and may result in product recalls. Therefore, thorough packaging inspection throughout the entire manufacturing and distribution process is essential.

On the 11th, Cognex announced the use case of the 'Insight Vision System', a deep learning-based machine vision solution used for packaging inspection automation, and an image-based barcode reader.

The Insight Vision System is a deep learning-based machine vision solution that integrates higher levels of automation into manufacturing, enabling companies to perform packaging inspection more efficiently. Cognex’s combination of machine vision, deep learning, barcode reading, and barcode verification technologies ensures that primary and secondary packaging is properly sealed, assembled without tampering, and ultimately defect-free throughout the manufacturing and distribution process, reducing unnecessary rework and preventing product recalls.

According to the Stericycle Expert Solutions Recall Index 2018, bacterial contamination was the leading cause of product recalls in the food and beverage industry. Therefore, packaging seal inspection, which determines the integrity of the packaging seal, plays a very important role in preventing recalls by ensuring the quality of the final product.

However, it is not easy to inspect due to changes in the appearance of defects that can occur during product sealing, and it is difficult to adjust for changes that occur during inspection or to classify or quantify the reasons for problems with the seal.

This deep learning-based machine vision solution highlights issues such as changes in particle size, contrast variations, and random defects that can alter the appearance of the packaging in real time, and supports operators and machines to classify the issues, enabling reliable identification of numerous issues that can affect product seals, including foreign substances, seals with empty spaces, and contamination.

Label quality inspection is a very important process for smooth product history management. However, manufacturers were having difficulty with inspection because there were many defects that were difficult to detect on labels due to the direction of the factory conveyor belt or the curved surface of the product.

To this end, Cognex uses machine vision systems with built-in defect detection technology to inspect labels for clean, accurate application, without wrinkles, bubbles, tears or other errors, while its In-Sight vision systems with feature extraction technology use lighting and software algorithms to create high-contrast images that enhance 3D features on products to capture all errors and defects, even torn or warped labels.

In addition, labels indicating common allergens (substances that cause allergies) such as peanuts, soybeans, milk, eggs, shellfish, tree nuts, and wheat must be attached to the product and the entire distribution process must be tracked. Cognex Insight vision systems help prevent this damage by using onboard pattern matching technology to thoroughly inspect packages, containers, and other items for the presence of allergen labels, as well as for the correct labeling and clear printing.

Additionally, the Insight Vision System, equipped with OCRMax technology that can manage and read any printed font, can detect the presence of codes and verify the accuracy of data. Even when codes are written on curved, reflective, or transparent surfaces such as medical devices or pharmaceuticals and are very difficult to read, it can detect the presence of product date and lot codes and check that serial numbers and characters are correct.

“Cognex accurately identified the needs of customers who wanted faster and more powerful machine vision systems for packaging, a key element that ensures the final quality of a product, and is providing an automated inspection system that combines deep learning with machine vision to meet even the most demanding customer requirements,” said Moon Eung-jin, CEO of Cognex Korea. “Companies can secure consistent product quality with minimal operator intervention, maximizing productivity and investment effectiveness while also improving brand reputation.”