Validation and classification of assembly defects in consumer electronic motherboards

Abstract

The assembly process in the electronics industry involves integrating various electronic boards and components to ensure proper functionality and communication. During this stage, defects such as missing, floating, and disconnected objects may arise. Failure to detect these can lead to product malfunctions, crashes, or inoperability. This work highlights challenges in a real manufacturing environment, specifically during the validation phase of consumer electronic production. Addressing these issues is crucial and useful, as similar defects can occur in similar consumer electronics manufacturing settings. In this work, we present a solution that integrated a multi-class classifier trained to detect missing and disconnected objects which achieves an average accuracy of 99%.