Vaccine Kit Assembly Verification
Confirm proper assembly of vaccination and medical kits, even as their components vary
Vaccination kits can contain a variety of parts, and if one is missing, damaged, or incorrect, the entire kit can be useless—in addition to being a significant recall risk. Since vaccine kits are typically still manually assembled, they are labor intensive and prone to error. Manufacturers are turning to automated robotic pick, place, and pack processes that quickly and more accurately assemble and inspect these kits, optimizing quality and avoiding expensive recalls.
Since vials, ampoules, and prefilled syringes are made of glass or rigid plastic, they must be packaged snugly to prevent breakage or cracking during shipment—but tight packaging can also result in damage, either to the container or the packaging, if there is any misalignment or other error.
Conventional machine vision requires a variety of algorithms, with rules programmed by a vision expert, to recognize the wide range of possible products and their placement. Changes in kit composition necessitates reprogramming and can be costly and take time.
Cognex Deep Learning is extremely effective at solving this type of application. The part location tool easily handles the many complexities of vaccine kit inspection such as parts placed at different orientations, overlapping parts, missing parts, or different kit combinations. It is easy to train a deep learning system to identify kit components at a wide variety of angles, and to identify and distinguish new parts, even ones with similar appearances.
For any kit, the locate tool can build a component library from a training data set. The desired layout can be established, then updated and changed for different production runs with different components.