A Novel Label Detection Method Based on Label Design Documents for Dynamic Industrial Environments

Abstract

In the manufacturing sector, labeling constitutes a crucial step prior to product distribution. Product labels play a vital role in providing customers with essential information about the products. Thus, ensuring the accuracy and quality of labeling is a critical component of overall product quality assurance. Given the necessity of verifying labeling accuracy, it is imperative to have reliable methods for testing labeling quality.In production lines where products are frequently changed, conventional label detection algorithms often fall short. To address this issue, a novel label detection method based on the label design document has been proposed. This method facilitates the automatic generation of classification rules derived from label design files, enabling the identification of unique features. Consequently, it allows for the precise detection of labeling errors, ensuring that products are labeled correctly.