Inspecting Adhesive Labels on Cosmetics
Check applied package labels for bubbles, creases, and other defects
Graphical programming environment for deep learning-based industrial image analysis
Many personal care products, such as lotions, ointments, and creams, are sold in containers like bottles and jars with curved surfaces. After labels with branding and product information are applied to the containers, they are inspected to ensure there are no bubbles, tears, wrinkles, or other defects.
The labels are curved and presented to the camera at many different angles. There is a wide variety of possible defects, including printing errors, which can appear anywhere on the label. The printed label can also have a confusing patterned or colored background, which complicates defect detection. It is not feasible to program all the many possible defects into traditional rule-based machine vision, which can be easily confused by background patterns.
If not identified before proceeding to further packaging, containers with defective labels will be rejected during palletizing. If not detected at that point, the end retailer will send them back, and this will impact the retailer’s trust in the product.
Cognex AI-based vision systems with edge learning technology help improve product quality by detecting any label application anomalies, no matter the label and the angle at which it is presented to the vision system. If label design or container shape changes, the classification tool quickly retrains on a set of images of the new label or container and can be back on-line with minimal delay. The intuitive development environment enables rapid turnaround to minimize machine downtime and increase operational equipment effectiveness (OEE).
- Ensure high-quality labels
- Maintain visual brand value
- Minimize changeover time to increase OEE