Challenges
Previously, the AOI stations on the Printed Circuit Board Assembly (PCBA) line generated over 10,000 false positive inspection points each week, leading to additional manual inspections and increased operational costs. To address this issue, Lanner integrated its advanced computer vision algorithms with the Edge AI appliances.
Solutions
The edge AI appliances were deployed as inference server to enhance the AOI system. This setup enabled real-time identification of missing components and tombstone defects in PCBAs.
The results were notable: the system achieved over 96% accuracy in defect detection and reduced false detections by up to 80%. This deployment underscores the effectiveness of Lanner’s Edge AI appliances in enhancing the precision and efficiency of manufacturing processes, demonstrating their capability to drive significant improvements in operational performance through advanced AI solutions.
Edge AI Appliances
For this use case, Lanner has deployed the Edge AI appliances LEC-2290E and EAI-I731:
LEC-2290E: A NVIDIA-Certified industrial-grade computer featuring an Intel® Core™ i7 CPU and NVIDIA® A2 GPU for high-performance AI processing. It offers extensive I/O options, 2 removable HDD/SSD drive bays, and LTE/WiFi connectivity for seamless data management and network integration.
EAI-I731: Lanner’s latest Edge AI appliance, equipped with a 14th Gen Intel Core i Processor and supporting NVIDIA L4 or A2 GPUs. This appliance provides rich I/O ports, 2 removable HDD/SSD drive bays, and 5G, LTE, and WiFi connectivity, making it ideal for robust edge inference in industrial environments.
Conclusion
By leveraging Lanner’s Edge AI appliances and Chooch’s advanced computer vision algorithms, Lanner has significantly enhanced the accuracy and efficiency of its PCBA AOI system. This successful deployment highlights the transformative potential of Edge AI in manufacturing, enabling real-time, precise visual inspection and improving overall operational efficiency