Abstract
A multi-modal manufacturing build monitoring platform ingests time-stamped functional test station results and inspection imagery from multiple factory stations. The platform stores the data in a unified store linked by unit identifiers and provides build-level, station-level, and device-level monitoring views. Statistical drift detection flags shifts in test result distributions, including capability metric changes and distribution shape changes. Computer vision models analyze AOI/SMT and other inspection images to detect visual defects such as solder anomalies, component placement issues, and chip/package damage. A correlation engine links visual defect patterns with numerical test trends across stations and units and generates ranked failure modes and AI-generated root-cause hypotheses using historical build data and cross-station correlations. The platform automatically generates issue-tracking tickets that include structured failure descriptions, hypotheses, and references to relevant test records and images, optionally with language translation.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Anonymous, "Multi-Modal AI Platform for Integrated Manufacturing Build Monitoring and Automated Defect Analysis", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10722