Code reading on chicken parts packaging
Accurately decipher human-readable codes on flexible transparent food packaging
Graphical programming environment for deep learning-based industrial image analysis
Chicken parts, cuts of meat, and ground meat are packaged for sale in a variety of transparent plastic shrink and stretch wraps. These packages are identified by date, lot, and other identification information inkjet printed on the wrap itself.
Because the contents are visible through the wrap, which itself may be preprinted with brand logos and other standard design and information elements, the text code must be read against a complex and unpredictable background. The code itself is printed directly on flexible plastic wrap and so is often skewed, smeared, or otherwise defective.
Errors can lead to products being sold past the expiration date, spoilage, and misdelivery, with consequent safety risks and loss of revenue. The skews, smears, and other defects in the printed text, and the complex and unpredictable background against which it must be read makes it difficult to program conventional machine vision to read the code reliably.
Cognex AI-enabled OCR tools read printed alphanumeric codes on challenging flexible wrap, even against the varying background of various cuts of meat or chicken. The OCR tool comes with a pretrained font library, making it easy to set and deploy. The tool trains on a small set of images of skewed, angled, and deformed text and then finds and reads such text on flexible food packaging, no matter what the product underneath the wrapping.
Grocery stores and consumers receive the chicken parts or cuts of meat they are expecting, even for visually variable organic products.