Cosmetic defect detection
Capture defects on challenging food and beverage packaging surfaces
A breakthrough in complex inspection, part location, classification, and OCR
One of the biggest challenges when dealing with cosmetic, or surface, defects on food and beverage packages is that they are dynamic, often caused by the forming process. Typical defects like hits, scratches, or stains may be indiscernible during early production. These defects only become visible under specific lighting conditions later in the production process. While the cost of late detection can be painfully high, so can false rejects. This inspection is especially important in regulated industries where poor packaging quality can lead to recall events or customer complaints.
Conventional vision technology can often miss complex cosmetic packaging defects such as bubbles in labels, color degradation, scratches, cracks, overprint, and over-wrap or under-wrap issues. These types of unpredictable defects or variations are easy to discern by human inspectors, but very difficult to program with rule-based machine vision algorithms.
Cognex AI-based-based image analysis software detects cosmetic defects on surfaces as reliably as human inspectors, but with the speed of a computerized system. The defect detection tool catches defects on coarse material with standard illumination, even when image quality is poor, by forming a reliable model of the shape and texture based on training images. From here, it identifies deviations in the surface texture as anomalies and uses a classification tool to classify them as hits or scratches. Consistently delivering high-quality products to the market translates to happier customers and improved brand loyalty.