Printed Circuit Board Inspection: Fusion of Optical and X-ray Images (FOXi) for Electronic Components Classification

The effective and efficient quality, security and provenance assurance of microelectronics and electronic products is a persisting challenge for the industry. The impetus towards finding automatic solutions that create high confidence about the quality and origin of components assembled on Printed Circuit Boards (PCBs) is high. A novel approach is proposed for microelectronic component detection and classification with high accuracy and confidence. A dedicated multi-source data acquisition, fusion and interpretation, are suggested to reduce uncertainty regarding concurrency, reliability, a vailability, and physical/logical resolution. Increasing the quality of data and subsequently generating the infrastructure for knowledge creation is the main output of this research. A central feature of the proposed solution is the integration and fusion of images from surface and penetrating imaging systems, set specifically to generate the necessary raw data. Experiments reported here demonstrate the merits of the accomplished classification, learning and identification of constituent components through the proposed fusion algorithm
¹
, The accuracy of the analytics from the fusion of optical and X-ray algorithms has improved by 5.3 %, compared to optics only. All other sensitivities were also significantly improved; for example, the Chip sensitivity in optical and X-ray images was 89.4% and 85.8%, respectively. With FOXi the metric increased to 94%.