Inline Color Vision for Specific Electroplating Defect Identification

Electroplating quality is established through a number of factors, where the visual appearance tends to be the immediate indicator of a surface quality defect. In high-volume connector manufacture, electroplating is a continuous production step, yet quality inspection is an offline, operator-dependent process. In this paper, the objective was to determine whether a color vision system could identify the variation in an electroplated deposit color and attribute this change to a particular defect. Design of experiments (DOE) was used to indicate the correct factor adjustments required to obtain the maximum color difference recorded between deposits. The vision system was then subjected to varying degrees of a burnt defective deposit. The outputs were analyzed using statistical tools to establish whether transitions from a good to a burnt deposit could be identified. The research has been implemented into an industrial application, and vital elements such as effective user feedback and defect traceability have been addressed. To establish an effective solution in practice required dedicated software to be designed and developed due to the challenges of both speed and volume of data. The benefits are seen through the combination of (1) the removal of a subjective visual inspection and (2) an increased detection process that is robust, repeatable, and reproducible.