Identification Feature Design for Homogenization Product Based on Characteristic Difference

Abstract

Aiming at the serious homogenization problems between the products with the same model and batches, and the difficulty identification due to wear, pollution and other factors which led small characteristic differences between cutters failure to identify, an identification feature design method of homogenization product to prevent tracking failure was presented. The feature enhanced direct parts marking (DPM) was analyzed, Data Matrix (DM) codes feature design based on marking process and affine transformation was proposed. Feature differences between individuals were enhanced from the source, DM code texture feature design based on Gray-level Co-occurrence (GLCM) was proposed and DM code texture library was established in product manufacturing process. By using the method of Hamming distance and Euclidean distance, the similarity between the target characteristics and DM code marking feature, manufacturing process texture library was matched when product identification was failed to read. The experimental results showed that the proposed method could identify the product failure recovery or remedy better.