Automated Glass Vial Counting
Accurately count transparent glass containers in large trays
Many liquid luxury consumer products such as oils, lotions, and perfumes are packaged in glass vials, bottles, or jars, for both chemical stability and elegant appearance. The container simultaneously preserves the premium product and shows it off.
Filled containers come densely packed on trays and must be counted before moving on to secondary packaging. The transparent and reflective glass containers are in contact with each other. The wide field of view necessary to view the entire tray shows containers from different angles in different parts of the image. All of this makes it difficult for conventional machine vision to count the containers accurately.
Accurate counting is important for process control, inventory control, and product reconciliation, to minimize any possible loss of high-value product.
Cognex Deep Learning is an ideal solution for automated non-contact counting. The part location tool learns from a small set of images of filled trays and then identifies and locates each container even when the part appears differently to the camera at different angles and adapts to variations in light reflections from the surfaces and contents or refractions through the transparent glass.
Deep Learning eliminates the need for mechanical counting, and so does not need to move or touch any of the containers, while ensuring 100 percent product reconciliation and preventing reworks to correct miscounts.