Pack Packaging Defect Classification
Locate and classify defects on cigarette packages
Graphical programming environment for deep learning-based industrial image analysis
Cigarette packs are assembled on a high speed rotating wheel and occasional defects occur, such as tears or punctures. Traditionally, various defects are just treated as “no good” and the packages are either fixed or scrapped as the high-speed operation makes it hard for a human to identify and manually keep track of specific defects. Each type of defect may reflect a problem in the manufacturing process or the raw materials. However the absence of digital classification and results data makes it difficult to diagnose and fix the root cause of the problem in a timely manner. The manufacturer incurs higher scrap costs as well as time and money to identify and correct the source of the quality problem.
Cognex AI-based technology locates and classifies packaging defects, such as tears or punctures, on cigarette packs, which helps to identify potential production issues that affect quality.
Using the classify tool, an engineer trains the software in with both “good” images as well as “no good” images that represent the different kinds of defects expected. During runtime, the model identifies the various defects and keeps track of how many of each were identified. Based on the model developed during training, Cognex edge learning accurately classifies cigarette packages during runtime. Valuable defect-specific data from this solution enables quality improvement and operations teams to diagnose the problem and fix it, which in turn improves product quality, reduces scrap costs as well and improves overall equipment efficiency.