Improve AOI Performance Through Smart Visual Insight Solutions Collaboration
Authors: Wayne Zhang, Ziv Zhao, Ben Wu, Peng Tang,Yan Wang, Marie Cole, Li Qin Shen, Andrew Vogel Company: IBM Corporation Date Published: 9/22/2019
Abstract: With technology advancements, the EMS industry is seeing increased product complexity from component miniaturization, component IO and design density increases. This trend is particularly challenging for the Automatic Optical Inspection (AOI) process in terms of inspection algorithm optimization and the operator training for the review of AOI rejects, in order to assure the product quality and production efficiency. The technology evolution in the Cloud, AI, IoT and Smart manufacturing has provided an opportunity to further improve the performance of AOI. This paper will present the POC (Proof Of Concept) trial of such a solution to improve AOI performance, including the edge computing from the AOI output, and the AI center in the cloud to process the AOI images and improve the classification of the image output. The paper will also cover a proposal to overcome the challenges observed in the development of this solution, which will include an industrial wide collaboration and ecosystem development.